Phoneme Boundary Research Articles

Bimodal automatic speech segmentation based on audio and visual information fusion

Bimodal automatic speech segmentation using visual information together with audio data is introduced. The accuracy of automatic segmentation directly affects the quality of speech processing systems using the segmented database. The collaboration of audio and visual data results in lower average absolute boundary error between the manual segmentation and automatic segmentation results. The information from two modalities are fused at the feature level and used in a HMM based speech segmentation system. A Turkish audiovisual speech database has been prepared and used in the experiments. The average absolute boundary error decreases up to 18% by using different audiovisual feature vectors. The benefits of incorporating visual information are discussed for different phoneme boundary types. Each audiovisual feature vector results in a different performance at different types of phoneme boundaries. The average absolute boundary error decreases by approximately 25% by using audiovisual feature vectors selectively for different boundary classes. Visual data is collected using an ordinary webcam. The proposed method is very convenient to be used in practice.

Phoneme Segmentation Based on Wavelet Spectra Analysis

A phoneme segmentation method based on the analysis of discrete wavelet transform spectra is described. The localization of phoneme boundaries is particularly useful in speech recognition. It enables one to use more accurate acoustic models since the length of phonemes provide more information for parametrization. Our method relies on the values of power envelopes and their first derivatives for six frequency subbands. Specific scenarios that are typical for phoneme boundaries are searched for. Discrete times with such events are noted and graded using a distribution-like event function, which represent the change of the energy distribution in the frequency domain. The exact definition of this method is described in the paper. The final decision on localization of boundaries is taken by analysis of the event function. Boundaries are, therefore, extracted using information from all subbands. The method was developed on a small set of Polish hand segmented words and tested on another large corpus containing 16 425 utterances. A recall and precision measure specifically designed to measure the quality of speech segmentation was adapted by using fuzzy sets. From this, results with F-score equal to 72.49% were obtained.

Evidence for atypical categorical speech perception in Williams syndrome

Although language processing has been described as being a relatively ‘spared’ cognitive function in individuals presenting with Williams syndrome (7q11.23 microdeletion; WS), some studies suggest possible alterations at the level of input speech processing. We explored categorical speech perception and non-speech auditory perception in six participants with WS, as well as in chronological age-matched or reading age-matched control groups. Categorical speech perception was explored for the b-d speech sound continuum, where speech sounds vary in spectral acoustic features, and for the d-t speech sound continuum, where speech sounds vary in temporal acoustic features. Non-speech perception was explored for sine-wave analogues of the b-d continuum, which are not identified as speech stimuli. We observed a significantly increased ability in WS participants to detect subtle acoustic changes between adjacent stimuli of the b-d or the d-t continuum outside the phoneme boundary, when control participants showed absence of discrimination. This is the first study to provide evidence for an atypical sensitivity towards subtle acoustic variations during speech and non-speech auditory analysis in WS. Implications for phonological processing and reading acquisition are discussed.

Search Key Identification in a Spoken Query using Isolated Keyword Recognition

article presents a novel technique for the recognition of isolated keywords from spoken search queries. Recognition of the isolated keywords from spoken search queries may be considered as the first step towards the development of a speech-operated keyword-based searching technique. A database of 300 spoken search queries from Assamese language, a major Indian language mostly spoken by the people of north east India, has been created. The system developed during the study has been tested and evaluated with the above mentioned database. In the present study, Mel Frequency Cepstral Coefficient (MFCC) has been used as the feature vector and Multilayer Perceptron (MLP) to identify the phoneme boundaries as well as for recognition of the phonemes. Viterbi search technique has been used to identify the keywords from the sequence of phonemes generated by the phoneme recognizer. A recognition accuracy of 74.67% has been achieved in the present study.

On the Importance of Transition Regions for Automatic Speaker Recognition

This letter describes the importance of transition regions, e.g. at phoneme boundaries, for automatic speaker recognition compared with using steady-state regions. Experimental results of automatic speaker identification tasks confirm that transition regions include the most speaker distinctive features. A possible reason for obtaining such results is described in view of articulation, in particular, the degree of freedom of articulators. These results are expected to provide useful information in designing an efficient automatic speaker recognition system.

The influence of talker gender, formant-frequency scaling, and presumed sources of variance on the perception of voiceless fricatives in American English.

We examined listeners’ identification of sigh-shy and sigh-thigh stimuli created by appending /s/-/∫/ and /s/-/θ/ continua to tokens of /ɑɪ/ excised from the natural productions of sigh spoken by one man and one woman. The values of the formants for the vocalic bases were scaled to give the illusion of differences in talkers’ overall stature. Five levels of scaling were created for both the man and woman’s productions. For the /s/-/∫/ continuum, listeners perceived more /∫/ with women’s voices than with men’s voices but, contrary to earlier research, showed gradient effects of formant scaling on phoneme boundaries for both the male and female tokens. For the /s/-/θ/ continuum, more /s/ tokens were identified for women’s voices than for men’s voices. Gradient effects of formant-frequency scaling were found for both men’s and women’s voices, and these were much smaller than those for /s/-/∫/. Listeners who were led to believe that the scaling reflected differences in talkers’ age height showed a bigger influence of scaling on /s/-/θ/ perception than did listeners who were told they reflected differences in height. Together, these results show that perceived gender, vocal-tract normalization, and social stereotypes influence fricative identification.

Left premotor cortex and allophonic speech perception in dyslexia: A PET study

Disorders of categorical perception has been put forward as a new account of phonological deficit in dyslexia (Serniclaes, W., Sprenger-Charolles, L., Carre, R. and Demonet, J.F., 2001. Perceptual discrimination of speech sounds in developmental dyslexia. J. Speech Lang. Hear. Res. 44, 384–399.) so that dyslexic subjects tend to discriminate phoneme instances within a given phonemic category rather than between categories, possibly witnessing the persistence of phonemic boundaries of ‘allophones’ that may be relevant to other languages although not to one's mother tongue (Serniclaes, W., Van Heghe, S., Mousty, P., Carre, R. and Sprenger-Charolles, L., 2004. Allophonic mode of speech perception in dyslexia. J. Exp. Child Psychol. 87, 336–361.). The brain correlates of within- and between-category discrimination were explored using a /ba/–/da/ phonetic continuum and H 2 15O PET in 14 dyslexic and 16 control adult readers; subjects discriminated a set of stimuli pairs, first in a ‘naïve’ (acoustic) condition and, after debriefing about the stimuli identity, in a speech (phonemic) condition (Dufor, O., Serniclaes, W., Sprenger-Charolles, L. and Demonet, J.F., 2007. Top-down processes during auditory phoneme categorization in dyslexia: a PET study. NeuroImage 34, 1692–1707.). While discrimination of ‘between’ pairs improved in all subjects following debriefing, ‘within’ stimuli yielded variable performance; some subjects kept discriminating them, while best categorizers judged them identical. Correlation analyses between acoustic-to-speech changes in brain activity and in ‘within’-pair discrimination, and between control and dyslexic groups, revealed a criss-crossed correlation pattern in the left BA6 so that the higher the activity the better the categorization in control subjects whereas the higher the activity the more increased ‘within’ discrimination in dyslexic subjects. Therefore, in average readers, enhanced activity in the left BA6 likely contributes to optimizing phoneme categorization via refined speech motor coding. In dyslexic subjects showing sensitivity to ‘within’-category cues, activity enhancement in this region might suggest the persistence of motor coding for allophonic representations of speech.

Distinctive Phonetic Feature (DPF) Extraction Based on MLNs and Inhibition/Enhancement Network

This paper describes a distinctive phonetic feature (DPF) extraction method for use in a phoneme recognition system; our method has a low computation cost. This method comprises three stages. The first stage uses two multilayer neural networks (MLNs): MLNLF-DPF, which maps continuous acoustic features, or local features (LFs), onto discrete DPF features, and MLNDyn, which constrains the DPF context at the phoneme boundaries. The second stage incorporates inhibition/enhancement (In/En) functionalities to discriminate whether the DPF dynamic patterns of trajectories are convex or concave, where convex patterns are enhanced and concave patterns are inhibited. The third stage decorrelates the DPF vectors using the Gram-Schmidt orthogonalization procedure before feeding them into a hidden Markov model (HMM)-based classifier. In an experiment on Japanese Newspaper Article Sentences (JNAS) utterances, the proposed feature extractor, which incorporates two MLNs and an In/En network, was found to provide a higher phoneme correct rate with fewer mixture components in the HMMs.

Two languages in one brain - slower phonological processing

Event Abstract Back to Event Two languages in one brain - slower phonological processing Maija Peltola1, 2*, H. Tamminen1, 2, H. Lehtola1, 2, T. Kujala3 and R. Näätänen3, 4, 5 1 Department of Phonetics, University of Turku, Finland 2 Centre for Cognitive Neuroscience, University of Turku, Finland 3 Department of Psychology, University of Helsinki, Finland 4 Department of Psychology, University of Tartu, Estonia 5 Center of Functionally Integrative Neuroscience, University of Aarhus, Denmark Phoneme perception is known to be language-specific so that mother tongue contrasts are discriminated easily and preattentively on the basis of long term memory traces. Behavioral and psychophysiological studies have established that discrimination sensitivity is at its peak at the native language phoneme boundaries and lowest near the native speech sound category prototypes. Bilinguals have a large amount of phonological categories that need to be kept apart. The aim of the present study was to determine whether monolinguals and bilinguals perceive speech sounds similarly or whether the two systems in bilinguals interact in a manner that one native language may be affected by the other. Two groups participated in the experiment: Group 1 consisted of 10 native speakers of Finnish and Group 2 of 12 balanced Finnish-Swedish bilinguals. Both groups were tested by the same Finnish speaking researcher. The stimuli were individually chosen on the basis of a prior identification experiment, which covered the rounded closed vowel continuum ranging form /y/ to /u/. The stimulus pair crossed the Finnish phoneme boundary /y/-/u/, but it was within the Swedish fronted /u/ vowel. We analyzed both the MMN peak latency (150-300 ms) and amplitude (two consecutive time windows 180-230 ms and 230-280 ms). The latency analysis revealed a significant effect of group resulting from a longer MMN latency in the bilingual than monolingual group. Consistently, the amplitude analysis revealed a significant interaction between the time window and group, again showing the lateness of the bilingual response. In addition, the amplitude analysis showed a group main effect suggesting a reduced MMN amplitude in the bilingual group. The groups differed significantly in their responses in the fist time window, where the monolinguals’ response was larger, but no group differences were found in the second time window. These results suggest slower preattentive processing in bilinguals than in monolinguals. This might be due to the more extensive intertwined phonological system from which items are accessed more slowly. The result showing a reduced MMN amplitude in bilinguals may be explained by interference from the other system, where the same contrast is non-phonemic, thus resulting in a small response. Conference: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications, Budapest, Hungary, 4 Apr - 7 Apr, 2009. Presentation Type: Poster Presentation Topic: Poster Presentations Citation: Peltola M, Tamminen H, Lehtola H, Kujala T and Näätänen R (2009). Two languages in one brain - slower phonological processing. Conference Abstract: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications. doi: 10.3389/conf.neuro.09.2009.05.149 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 26 Mar 2009; Published Online: 26 Mar 2009. * Correspondence: Maija Peltola, Department of Phonetics, University of Turku, Turku, Finland, maipie@utu.fi Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Google Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Google Scholar Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen PubMed Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

Different kinds of bilinguals - different kinds of brains

Event Abstract Back to Event Different kinds of bilinguals - different kinds of brains Maija Peltola1, 2*, H. Tamminen1, 2, H. Lehtola1, 2, T. Kujala3 and R. Näätänen3, 4, 5 1 Department of Phonetics, University of Turku, Finland 2 Centre for Cognitive Neuroscience, University of Turku, Finland 3 Cognitive Brain Research Unit, Department of Psychology, University of Helsinki, Finland 4 Department of Psychology, University of Tartu, Estonia 5 Center of Functionally Integrative Neuroscience, University of Aarhus, Denmark Earlier research on bilingual speech processing has suggested that bilinguals may have two separate linguistic systems, which preattentively respond more strongly to the language used in the immediate vicinity. Also, these two systems seem to have different neural representation areas. However, some studies have indicated contradictory results by showing that bilinguals respond equally to phonological contrasts from both languages and that the representations might also be identical. These discrepancies may be related to the types of bilinguals included in different studies. The aim of this study was to determine how different kinds of bilinguals may differ in their preattentive phonological discrimination. Two groups participated in the experiment: Group 1 consisted of 12 balanced Finnish-Swedish bilinguals, who had acquired both their languages from birth, while the 9 learner bilinguals in Group 2 were advanced Finnish University students of Swedish. Half of the subjects participated first so that only Swedish was used with the Swedish speaking researcher, while the other half was first tested in the Finnish contexts with a Finnish researcher. The stimuli were selected individually for each subject on the basis of an identification experiment which covered the rounded closed vowel continuum from /y/ to /u/. The continuum is divided into two categories in Finnish, but Swedish has the additional fronted /u/ phoneme. We selected a vowel pair, which crossed the phoneme boundary in Finnish, but was within the fronted Swedish /u/. Consequently, it is possible to see whether the same acoustic difference results in different kinds of responses in two types of bilinguals. Our results showed a significant interaction between linguistic context and group. Further analysis revealed a significant effect of the context language in the learner bilinguals while the balanced bilinguals showed no such difference. Thus, the learner group showed a larger response for the same vowel pair, when the context made it relevant, while the same pair elicited a smaller response when it was non-significant in the light of the context language. In contrast, the balanced bilinguals showed no such dissociation, since similar responses were elicited by the same pair irrespective of contexts. This result shows that balanced bilinguals perceive speech sounds automatically from both languages, while learner bilinguals perceive phonemes context-specifically. This suggests that balanced bilinguals might have one intertwined system, which contains the categories of both languages, while learner bilinguals have two separate systems, which can be accessed independently. Conference: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications, Budapest, Hungary, 4 Apr - 7 Apr, 2009. Presentation Type: Poster Presentation Topic: Poster Presentations Citation: Peltola M, Tamminen H, Lehtola H, Kujala T and Näätänen R (2009). Different kinds of bilinguals - different kinds of brains. Conference Abstract: MMN 09 Fifth Conference on Mismatch Negativity (MMN) and its Clinical and Scientific Applications. doi: 10.3389/conf.neuro.09.2009.05.150 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 26 Mar 2009; Published Online: 26 Mar 2009. * Correspondence: Maija Peltola, Department of Phonetics, University of Turku, Turku, Finland, maipie@utu.fi Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Google Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Google Scholar Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen PubMed Maija Peltola H. Tamminen H. Lehtola T. Kujala R. Näätänen Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

Testing the double phonemic boundary in bilinguals

It is widely known that language influences the way speech sounds are categorized. However, categorization of speech sounds by bilinguals is not well understood. There is evidence that bilinguals have different category boundaries than monolinguals, and there is evidence suggesting that bilinguals’ phonemic boundaries can shift with language context. This phenomenon has been referred as the double phonemic boundary. In this investigation, the double phonemic boundary is tested in Spanish-English bilinguals ( N = 18) and English monolinguals ( N = 16). Participants were asked to categorize speech stimuli from a continuum ranging from /ga/ to /ka/ in two language contexts. The results showed phonemic boundary shifts in bilinguals and monolinguals which did not differ across language contexts. However, the magnitude of the phoneme boundary shift was significantly correlated with the level of confidence in using English and Spanish (reading, writing, speaking, and comprehension) for bilinguals, but not for monolinguals. The challenges of testing the double phonemic boundary are discussed, along with the limitations of the methodology used in this study.

Perceptual development of phoneme contrasts: How sensitivity changes along acoustic dimensions that contrast phoneme categories

Listeners discriminate acoustic differences between phoneme categories at a higher level than similarly sized differences within phoneme categories. The question this paper aims to answer is how this pattern in perceptual sensitivity develops along an acoustic dimension that contrasts two non-native speech sounds: through acquired distinctiveness, through acquired similarity, or through a combination of the two. A pretest-training-post-test experiment was designed to study perceptual development directly, i.e., by including (i) a discrimination task to measure perceptual sensitivity, (ii) a transfer test to ensure language learning instead of stimulus learning, and (iii) a control group to exclude task repetition as an explanation of improvement. It is shown that the typical peak in perceptual sensitivity near a phoneme boundary that native listeners show is not found in relatively inexperienced language learners, despite their ability to classify a continuum in a nativelike way after short laboratory training. Experiment II indicates that a discrimination peak may be achieved by language learners, but only after much more language experience than short-term laboratory training can offer. Furthermore, reasons are given why classification improvement in the laboratory should not be taken as evidence for (i) increased discrimination of the newly learned phonemes and (ii) learning of phoneme representations.

Visual speech contributes to phonetic learning in 6-month-old infants

Previous research has shown that infants match vowel sounds to facial displays of vowel articulation [Kuhl, P. K., & Meltzoff, A. N. (1982). The bimodal perception of speech in infancy. Science, 218, 1138–1141; Patterson, M. L., & Werker, J. F. (1999). Matching phonetic information in lips and voice is robust in 4.5-month-old infants. Infant Behaviour & Development, 22, 237–247], and integrate seen and heard speech sounds [Rosenblum, L. D., Schmuckler, M. A., & Johnson, J. A. (1997). The McGurk effect in infants. Perception & Psychophysics, 59, 347–357; Burnham, D., & Dodd, B. (2004). Auditory-visual speech integration by prelinguistic infants: Perception of an emergent consonant in the McGurk effect. Developmental Psychobiology, 45, 204–220]. However, the role of visual speech in language development remains unknown. Our aim was to determine whether seen articulations enhance phoneme discrimination, thereby playing a role in phonetic category learning. We exposed 6-month-old infants to speech sounds from a restricted range of a continuum between /ba/ and /da/, following a unimodal frequency distribution. Synchronously with these speech sounds, one group of infants (the two-category group) saw a visual articulation of a canonical /ba/ or /da/, with the two alternative visual articulations, /ba/ and /da/, being presented according to whether the auditory token was on the /ba/ or /da/ side of the midpoint of the continuum. Infants in a second (one-category) group were presented with the same unimodal distribution of speech sounds, but every token for any particular infant was always paired with the same syllable, either a visual /ba/ or a visual /da/. A stimulus-alternation preference procedure following the exposure revealed that infants in the former, and not in the latter, group discriminated the /ba/–/da/ contrast. These results not only show that visual information about speech articulation enhances phoneme discrimination, but also that it may contribute to the learning of phoneme boundaries in infancy.

The Use of Overlapped Sub-Bands in Multi-Band, Multi-SNR, Multi-Path Recognition of Noisy Word Utterances

A solution to the problem of improving robustness to noise in automatic speech recognition is presented in the framework of multi-band, multi-SNR, and multi-path approaches. In our word recognizer, the whole frequency band is divided into seven-overlapped sub-bands, and then sub-band noisy phoneme HMMs are trained on speech data mixed with the filtered white Gaussian noise at multiple SNRs. The acoustic model of a word is built as a set of concatenations of clean and noisy sub-band phoneme HMMs arranged in parallel. A Viterbi decoder allows a search path to transit to another SNR condition at a phoneme boundary. The recognition scores of the sub-bands are then recombined to give the score for a word. Experiments show that the overlapped seven-band system yields the best performance under nonstationary ambient noises. It is also shown that the use of filtered white Gaussian noise is advantageous for training noisy phoneme HMMs.

Evaluation of auditory processing disorder and auditory efferent system functionality in adult dyslexics: towards a unification theory of auditory-language processing impairments

In a recent paper, Veuillet et al., (2007) suggested that some auditory processing mechanism could be impaired in children with dyslexia. They reported a link between children's ability to perceive phonemic boundaries and the physiological functionality of their medial olivocochlear system (MOC), an auditory efferent pathway functioning under central control. In the present experiment, we extended this observation by comparing speech-in-speech comprehension performances in a group of control participants (N=40) and a group of adults who had been diagnosed dyslexic as children (N=49). Confirming the idea that patients with dyslexia present auditory processing disorders (APD), we show that adult dyslexics exhibit greater difficulty in comprehending speech in noise. Data moreover suggest a link between speech-in-noise comprehension difficulty and the MOC functionality in these participants. More precisely, it appears that the absolute functionality of the left and right MOC bundles does not differ in dyslexic patients compared to controls. What appears to be differing is the functional asymmetry in the MOC functionality between both ears, normal readers showing a classical functional asymmetry, while dyslexic adults show an absence of, or reduced functional asymmetry between MOC bundles. These results will be discussed in the context of current models of APD and Dyslexia.

Associations Between Key Language-Related Measures in Typically Developing School-Age Children

Associations Between Key Language-Related Measures in Typically Developing School-Age Children

Auditory-visual speech perception in normal-hearing and cochlear-implant listeners

The present study evaluated auditory-visual speech perception in cochlear-implant users as well as normal-hearing and simulated-implant controls to delineate relative contributions of sensory experience and cues. Auditory-only, visual-only, or auditory-visual speech perception was examined in the context of categorical perception, in which an animated face mouthing ba, da, or ga was paired with synthesized phonemes from an 11-token auditory continuum. A three-alternative, forced-choice method was used to yield percent identification scores. Normal-hearing listeners showed sharp phoneme boundaries and strong reliance on the auditory cue, whereas actual and simulated implant listeners showed much weaker categorical perception but stronger dependence on the visual cue. The implant users were able to integrate both congruent and incongruent acoustic and optical cues to derive relatively weak but significant auditory-visual integration. This auditory-visual integration was correlated with the duration of the implant experience but not the duration of deafness. Compared with the actual implant performance, acoustic simulations of the cochlear implant could predict the auditory-only performance but not the auditory-visual integration. These results suggest that both altered sensory experience and improvised acoustic cues contribute to the auditory-visual speech perception in cochlear-implant users.

Two-month-old infants at risk for dyslexia do not discriminate /bAk/ from /dAk/: A brain-mapping study

Dyslexics have problems with categorization of speech sounds, in particular when rapid temporal processing is involved such as in formant transitions of stop-consonants. Infants are already sensitive to such auditory features, but here we show that precursors of impaired categorization are already present in the brain responses of two-month-old infants at familial risk for dyslexia. Natural speech stimuli (/bAk/ and /dAk/), at either side of the phoneme boundary, induced multiple mismatch responses in control infants under pre-attentive and pre-cognitive conditions. Infants at-risk showed an attenuated early mismatch response and an absent late one, in addition to diminished cortical activity in the left hemisphere. These results are consistent with a temporal processing deficit in the infants at risk and may point to an early precursor of the disorder.

Automatic Phonetic Segmentation by Score Predictive Model for the Corpora of Mandarin Singing Voices

This paper proposes the concept of a score predictive model (SPM) that can refine the phoneme boundaries obtained by a hidden Markov model (HMM) and dynamic time warping (DTW) for a Mandarin singing voice corpus. An SPM is constructed by using support vector regression. It predicts the score of a phoneme boundary according to the boundary's 58-dimensional feature vector. The correctly identified boundaries of a singing corpus can then be used for corpus-based singing voice synthesis. Several experiments with different settings, including the use of different initial estimates, different acoustic features, and various regression approaches, were designed to verify the feasibility of the proposed approach. Experimental results demonstrate that the proposed SPM is able to effectively refine the results of the HMM and DTW.

Cortical categorization failure in 2-month-old infants at risk for dyslexia

Cortical auditory categorization was assessed in 2-month-old infants at genetic risk for dyslexia and controls. Manipulated /bAk/ and /dAk/ speech stimuli were used as frequently presented standards. The neuroelectric brain responses of the control infants were highly sensitive to the phoneme boundary that separated these stimuli, but the infants at risk showed no signs of cortical categorical perception. Cortical sources of the responses were predominantly found in the left hemisphere for the control infants, but mainly in the right hemisphere for the infants at risk. The results demonstrate that impaired categorical perception in dyslexia is already present in infants at risk for the disorder.