Phoneme Accuracy Research Articles

Impact of iPod Touch-Supported Repeated Reading on the English Oral Reading Fluency of L2 students with Specific Learning Difficulties

<span>In recent years the use of new technologies has been extensively explored in different aspects of language learning pedagogy. The objective of this research was to investigate the impact Repeated Reading activity, supported by iPod Touch could have on the English Oral Reading Fluency (ORF) of second language university students with Special Learning Difficulties (SpLD) at Cyprus University of Technology. As part of their university courses, students have two compulsory English courses. Due to their SpLD and low level of language competence, the eight participants enrolled in the English programme for students with SpLD. This programme is based on the phonological approach and the research done in methods dealing with dyslexia (Shaywitz et al., 2004). After being introduced to the iPod-supported Repeated Reading activity, students worked independently for 8 weeks. They listened and replicated three recorded texts performed by native speakers, using Voice Memo. Texts were based on specific phonetic rules the students had to master. Students recorded their best performance of each text reading, using DropVox. Curriculum-Based Measurement, adapted by Rasinski (2004), was used to measure students’ automaticity (speed and accuracy), and an adapted version of Zutell and Rasinksi’s (1991) Multidimensional Framework to measure prosodic features of fluency. A phonemic accuracy scale was developed and used to assess students’ performance related to specific phonemes students had difficulty with. Data analysis revealed that the independent out-of-class use of Repeated Reading, supported by iPod Touch technology helped in increasing students’ automaticity, improving their prosodic features of fluency, including that of specific phonemes.</span>

Solving the Phoneme Conflict in Grapheme-to-Phoneme Conversion Using a Two-Stage Neural Network-Based Approach

SUMMARY To achieve high quality output speech synthesis systems, data-driven grapheme-to-phoneme (G2P) conversion is usually used to generate the phonetic transcription of out-of-vocabulary (OOV) words. To improve the performance of G2P conversion, this paper deals with the problem of conflicting phonemes, where an input grapheme can, in the same context, produce many possible output phonemes at the same time. To this end, we propose a two-stage neural network-based approach that converts the input text to phoneme sequences in the first stage and then predicts each output phoneme in the second stage using the phonemic information obtained. The first-stage neural network is fundamentally implemented as a many-to-many mapping model for automatic conversion of word to phoneme sequences, while the second stage uses a combination of the obtained phoneme sequences to predict the output phoneme corresponding to each input grapheme in a given word. We evaluate the performance of this approach using the American English words-based pronunciation dictionary known as the auto-aligned CMUDict corpus[1]. In terms of phoneme and word accuracy of the OOV words, on comparison with several proposed baseline approaches, the evaluation results show that our proposed approach improves on the previous one-stage neural network-based approach for G2P conversion. The results of comparison with another existing approach indicate that it provides higher phoneme accuracy but lower word accuracy on a general dataset, and slightly higher phoneme and word accuracy on a selection of words consisting of more than one phoneme

California Indian Languages by Victor Golla (review)

Reviewed by: California Indian Languages by Victor Golla Catherine A. Callaghan California Indian Languages. Victor Golla. Berkeley and Los Angeles: University of California Press, 2011. Pp. vii + 380. $57.50 (cloth). “Anyone who has paid the slightest attention to the aboriginal languages of California knows that there were a lot of them.” This is the lead sentence to page 1 of the introduction to this excellent and comprehensive book. Its scope is far greater than its title suggests. By “California,” Victor Golla means “the California area.” Geographically, this comprises those languages spoken from approximately 31° 30′ N in Baja California to 43° N in south-central Oregon—a region considerably larger than California, since it includes all those language families which may have had a toehold in the state, such as Takelma, usually considered an Oregon language. This was one of the most linguistically complex regions in the world, encompassing an area nearly a thousand miles long and sometimes more than two hundred miles wide where at least seventy-eight mutually unintelligible languages were spoken. This number represents nearly a third of the languages of North America north of Mesoamerica. Part 1, “Introduction: Defining California as a Sociolinguistic Area,” focuses on the consequences of the tribelet as the basic political unit. The tribelet consists of the chief’s village plus a few smaller settlements. Common language was sometimes a unifying factor between tribelets, but it was secondary since villages speaking the same language were occasionally enemies, as was the case among some of the Eel River Athabaskan tribelets. Part 2, “History of Study,” details the documentation of California Indian languages in full historical context from European contact to the present, including information on where surviving documents are housed, and parts of this section will make the linguistic scholar weep. The Jesuits were originally in charge of missions in Baja California. It was their policy to learn, record, and preach in local languages. Most of their extensive records were lost when King Carlos III of Spain expelled the order in 1767, with the result that we know little of the languages around the lower part of the peninsula except for a sketch of Waikuri, which was probably a language isolate. Other languages spoken nearby included Monqui and Pericu, but so little data on them have survived that we can say almost nothing about them. As far as the languages are concerned, things fared better in Alta California, where the Spanish missions were established by the Franciscans, some of whom ignored orders from the King of Spain and translated prayers and catechisms into aboriginal languages. A few went even further, most notably Fray Felipe Arroyo de la Cuesta, a missionary from 1808 to 1840 who wrote extensively on these languages. His crown jewel was a grammar and phrase book of Mutsun, the language of San Juan Bautista. A century later, J. P. Harrington was able to reelicit Arroyo’s surviving corpus from Ascención Solorsano, the last fluent Mutsun speaker. Father Geronimo Boscana, a missionary at San Juan Capistrano, wrote a lengthy scholarly treatise on the aboriginal religion and culture as well as material on the language. As we progress through time, additional historical figures connected with California languages come alive under Golla’s skilled pen. We learn that the Frenchman, Alphonse Pinart, who collected extensive vocabularies of Mission Indian languages during the nineteenth century, devoted much of his fortune to the project and died in poverty. We also learn that the naturalist Clinton Hart Merriam, who collected the Indian names for [End Page 295] plants and animals throughout a good part of California in the early twentieth century and listed them with their English and Latin identifications, had earlier helped found the American Ornithologists’ Union and had traveled on an expedition to Alaska with John Muir. Unfortunately, Merriam insisted on transcribing Indian words in an orthography based on that of Webster’s dictionary, to the dismay of contemporary linguists. However, his Miwok transcriptions were accurate enough to trigger my consultants’ memories for forgotten items. Consequently, I was able to list the words in my dictionaries with both phonemic accuracy and Merriam’s precise Latin definitions. The most significant of...

Distinct developmental profiles in typical speech acquisition

Three- to five-year-old children produce speech that is characterized by a high level of variability within and across individuals. This variability, which is manifest in speech movements, acoustics, and overt behaviors, can be input to subgroup discovery methods to identify cohesive subgroups of speakers or to reveal distinct developmental pathways or profiles. This investigation characterized three distinct groups of typically developing children and provided normative benchmarks for speech development. These speech development profiles, identified among 63 typically developing preschool-aged speakers (ages 36-59 mo), were derived from the children's performance on multiple measures. These profiles were obtained by submitting to a k-means cluster analysis of 72 measures that composed three levels of speech analysis: behavioral (e.g., task accuracy, percentage of consonants correct), acoustic (e.g., syllable duration, syllable stress), and kinematic (e.g., variability of movements of the upper lip, lower lip, and jaw). Two of the discovered group profiles were distinguished by measures of variability but not by phonemic accuracy; the third group of children was characterized by their relatively low phonemic accuracy but not by an increase in measures of variability. Analyses revealed that of the original 72 measures, 8 key measures were sufficient to best distinguish the 3 profile groups.

Articulatory Behaviors of Nonnative Speakers: Role of L2 Proficiency and Accent Modification

To sound native-like, nonnative speakers need to approximate the articulatory patterns of native speakers. When nonnative speakers deviate from native speakers' production patterns, it gives rise to a nonnative accent. This study examines how proficiency in a second language (L2) is related to processing skills in L2, as measured by the accuracy of consonant production. 10 monolingual speakers of American English, 10 Bengali-English bilinguals with high proficiency in English, and 10 Bengali-English bilinguals with low proficiency in English participated. The participants spoke 16 nonwords of four different syllable lengths. Productions were phonetically transcribed. Results suggested that nonnative speakers' L2 proficiency differentially interacted with overall phonemic accuracy, usage of appropriate place, manner and voicing features, position of consonants within target words, and syllable length. Implications of these findings are discussed in light of customized strategies for modification of nonnative accents.

Effects of Syllable Language Model on Distinctive Phonetic Features (DPFs) based Phoneme Recognition Performance

This paper presents a distinctive phonetic features (DPFs) based phoneme recognition method by incorporating syllable language models (LMs). The method comprises three stages. The first stage extracts three DPF vectors of 15 dimensions each from local features (LFs) of an input speech signal using three multilayer neural networks (MLNs). The second stage incorporates an Inhibition/Enhancement (In/En) network to obtain more categorical DPF movement and decorrelates the DPF vectors using the Gram-Schmidt orthogonalization procedure. Then, the third stage embeds acoustic models (AMs) and LMs of syllable-based subwords to output more precise phoneme strings. From the experiments, it is observed that the proposed method provides a higher phoneme correct rate as well as a tremendous improvement of phoneme accuracy. Moreover, it shows higher phoneme recognition performance at fewer mixture components in hidden Markov models (HMMs).

Analysis and Recognition of NAM Speech Using HMM Distances and Visual Information

Non-audible murmur (NAM) is an unvoiced speech signal that can be received through the body tissue with the use of special acoustic sensors (i.e., NAM microphones) attached behind the talker's ear. The authors had previously reported experimental results for NAM recognition using a stethoscopic and a silicon NAM microphone. Using a small amount of training data from a single speaker and adaptation approaches, 93.9% of word accuracy was achieved for a 20 k Japanese vocabulary dictation task. In this paper, further analysis of NAM speech is made using distance measures between hidden Markov models (HMMs). It has been shown that owing to the reduced spectral space of NAM speech, the HMM distances are also reduced when compared with those of normal speech. In the case of Japanese vowels and fricatives, the distance measures in NAM speech follow the same relative inter-phoneme relationship as that in normal speech without significant differences. However, significant differences have been found in the case of Japanese plosives. More specifically, in NAM speech, the distances between voiced/unvoiced consonant pairs articulated in the same place drastically decreased. As a result, the inter-phoneme relationship as compared to normal-speech changed significantly, causing a substantial decrease in the recognition accuracy. A speaker-dependent phoneme recognition experiment has been conducted, obtained 81.5% NAM phoneme correct, showing a relationship between HMM distance measures and phoneme accuracy. In a NAM microphone, body transmission and loss of lip radiation act as a low-pass filter. As a result, higher frequency components are attenuated in a NAM signal. Because of spectral reduction, NAM's unvoiced nature, and the type of articulation, NAM sounds become similar, causing a larger number of confusions when compared with normal speech. Yet many of those sounds are visually different on face/mouth/lips, and the integration of visual information increases their discrimination. As a result, recognition accuracy increases as well. In this article, the visual information extracted from the talkers' facial movements is fused with NAM speech. The experimental results reveal a relative improvement of 10.5% on average when fused NAM speech and facial information were used compared with using only NAM speech.

English Speech Acquisition in 3- to 5-Year-Old Children Learning Russian and English

English speech acquisition in Russian-English (RE) bilingual children was investigated, exploring the effects of Russian phonetic and phonological properties on English single-word productions. Russian has more complex consonants and clusters and a smaller vowel inventory than English. One hundred thirty-seven single-word samples were phonetically transcribed from 14 RE and 28 English-only (E) children, ages 3;3 (years;months) to 5;7. Language and age differences were compared descriptively for phonetic inventories. Multivariate analyses compared phoneme accuracy and error rates between the two language groups. RE children produced Russian-influenced phones in English, including palatalized consonants and trills, and demonstrated significantly higher rates of trill substitution, final devoicing, and vowel errors than E children, suggesting Russian language effects on English. RE and E children did not differ in their overall production complexity, with similar final consonant deletion and cluster reduction error rates, similar phonetic inventories by age, and similar levels of phonetic complexity. Both older language groups were more accurate than the younger language groups. We observed effects of Russian on English speech acquisition; however, there were similarities between the RE and E children that have not been reported in previous studies of speech acquisition in bilingual children. These findings underscore the importance of knowing the phonological properties of both languages of a bilingual child in assessment.

Cued Speech automatic recognition in normal-hearing and deaf subjects

This article discusses the automatic recognition of Cued Speech in French based on hidden Markov models (HMMs). Cued Speech is a visual mode which, by using hand shapes in different positions and in combination with lip patterns of speech, makes all the sounds of a spoken language clearly understandable to deaf people. The aim of Cued Speech is to overcome the problems of lipreading and thus enable deaf children and adults to understand spoken language completely. In the current study, the authors demonstrate that visible gestures are as discriminant as audible orofacial gestures. Phoneme recognition and isolated word recognition experiments have been conducted using data from a normal-hearing cuer. The results obtained were very promising, and the study has been extended by applying the proposed methods to a deaf cuer. The achieved results have not shown any significant differences compared to automatic Cued Speech recognition in a normal-hearing subject. In automatic recognition of Cued Speech, lip shape and gesture recognition are required. Moreover, the integration of the two modalities is of great importance. In this study, lip shape component is fused with hand component to realize Cued Speech recognition. Using concatenative feature fusion and multi-stream HMM decision fusion, vowel recognition, consonant recognition, and isolated word recognition experiments have been conducted. For vowel recognition, an 87.6% vowel accuracy was obtained showing a 61.3% relative improvement compared to the sole use of lip shape parameters. In the case of consonant recognition, a 78.9% accuracy was obtained showing a 56% relative improvement compared to the use of lip shape only. In addition to vowel and consonant recognition, a complete phoneme recognition experiment using concatenated feature vectors and Gaussian mixture model (GMM) discrimination was conducted, obtaining a 74.4% phoneme accuracy. Isolated word recognition experiments in both normal-hearing and deaf subjects were also conducted providing a word accuracy of 94.9% and 89%, respectively. The obtained results were compared with those obtained using audio signal, and comparable accuracies were observed.

Learning Lexicons from Spoken Utterances Based on Statistical Model Selection

This paper proposes a method for the unsupervised learning of lexicons from pairs of a spoken utterance and an object as its meaning without any a priori linguistic knowledge other than a phoneme acoustic model. In order to obtain a lexicon, a statistical model of the joint probability of a spoken utterance and an object is learned based on the minimum description length principle. This model consists of a list of word phoneme sequences and three statistical models: the phoneme acoustic model, a word-bigram model, and a word meaning model. Experimental results show that the method can acquire acoustically, grammatically and semantically appropriate words with about 85% phoneme accuracy. Index Terms: Lexical learning, language acquisition, model selection.

Phonological patterns in Mandarin–English bilingual children

Adele Miccio recognized the paucity of information on the phonological development of children from diverse linguistic and cultural backgrounds, and emphasized the need to apply advances in bilingual phonological research toward an appropriate phonological measure for bilingual children. In the spirit of her pioneering work, the present study investigated both Mandarin and English phonological patterns in typically-developing 5-year-old bilingual children in an English-immersion programme in Taiwan. Consonant and vowel accuracy, number and types of phonological processes, and Mandarin-influenced English patterns were assessed on a single-word assessment in each language. Results indicated comparable levels of phoneme accuracy and similar rates and types of phonological processes for bilinguals and their monolingual counterparts. A number of English phonological processes for bilinguals, however, suggested a possible Mandarin influence. The present results reiterate Dr Miccio's call for interdisciplinary collaboration to enhance one's understanding of bilingual language development, to advance successful intervention for bilingual children.

The relationship between articulatory control and improved phonemic accuracy in childhood apraxia of speech: A longitudinal case study

Jaw movement patterns were examined longitudinally in a 3-year-old male with childhood apraxia of speech (CAS) and compared with a typically developing control group. The child with CAS was followed for 8 months, until he began accurately and consistently producing the bilabial phonemes /p/, /b/, and /m/. A movement tracking system was used to study jaw duration, displacement, velocity, and stability. A transcription analysis determined the percentage of phoneme errors and consistency. Results showed phoneme-specific changes which included increases in jaw velocity and stability over time, as well as decreases in duration. Kinematic parameters became more similar to patterns seen in the controls during final sessions where tokens were produced most accurately and consistently. Closing velocity and stability, however, were the only measures to fall within a 95% confidence interval established for the controls across all three target phonemes. These findings suggest that motor processes may differ between children with CAS and their typically developing peers.

Unintended imitation in nonword repetition

Verbal repetition is conventionally considered to require motor-reproduction of only the phonologically relevant content of a perceived linguistic stimulus, while imitation of incidental acoustic properties of the stimulus is not an explicit part of this task. Exemplar-based theories of speech processing, however, would predict that imitation beyond linguistic reproduction may occur in word repetition. Five experiments were conducted in which verbal audio-motor translations had to be performed under different conditions. Nonwords varying in phonemic content, in vocal pitch ( F 0), and in speaking style (schwa-syllable expression) were presented. We experimentally varied the factors response delay (repetition vs. shadowing), intention-to-repeat (repetition vs. pseudo-naming), and phonological load (repetition vs. transformation). The responses of ten healthy participants were examined for phonemic accuracy and for traces of para-phonological imitation. Two aphasic patients with phonological impairments were also included, to find out if lesions to left anterior or posterior perisylvian cortex interfere with imitation. In the healthy participants, significant imitation of both F 0 and phonetic style was observed, with markedly stronger effects for the latter. Strong imitation was also found in an aphasic patient with a lesion to left anterior perisylvian cortex, whereas almost no imitation occurred in a patient with a lesion to the posterior language area. The degree of unintended imitation was modulated by each of the three independent factors introduced here. The results are discussed on the background of cognitive and neurolinguistic theories of imitation.

A hybrid approach to bilingual text-to-phoneme mapping

In this paper, we address the problem of bilingual text-to-phoneme (TTP) mapping in which the phonetic transcription of isolated written words must be found. In general, in the bilingual/multilingual TTP mapping for isolated words, two processing steps are applied to each input word. The language of each word is first identified and then the letters of the word are translated into their phonetic transcriptions according to the recognized language. We use the multilayer perceptron (MLP) neural network for the letter to phoneme conversion task and a hybrid approach composed of a MLP and a decision rule system for the language recognition task. We introduce a new bilingual TTP mapping system and we provide an analysis of the influence of several different factors on its phoneme accuracy. .

English Speech Sound Development in Preschool-Aged Children From Bilingual English–Spanish Environments

English speech acquisition by typically developing 3- to 4-year-old children with monolingual English was compared to English speech acquisition by typically developing 3- to 4-year-old children with bilingual English-Spanish backgrounds. We predicted that exposure to Spanish would not affect the English phonetic inventory but would increase error frequency and type in bilingual children. Single-word speech samples were collected from 33 children. Phonetically transcribed samples for the 3 groups (monolingual English children, English-Spanish bilingual children who were predominantly exposed to English, and English-Spanish bilingual children with relatively equal exposure to English and Spanish) were compared at 2 time points and for change over time for phonetic inventory, phoneme accuracy, and error pattern frequencies. Children demonstrated similar phonetic inventories. Some bilingual children produced Spanish phonemes in their English and produced few consonant cluster sequences. Bilingual children with relatively equal exposure to English and Spanish averaged more errors than did bilingual children who were predominantly exposed to English. Both bilingual groups showed higher error rates than English-only children overall, particularly for syllable-level error patterns. All language groups decreased in some error patterns, although the ones that decreased were not always the same across language groups. Some group differences of error patterns and accuracy were significant. Vowel error rates did not differ by language group. Exposure to English and Spanish may result in a higher English error rate in typically developing bilinguals, including the application of Spanish phonological properties to English. Slightly higher error rates are likely typical for bilingual preschool-aged children. Change over time at these time points for all 3 groups was similar, suggesting that all will reach an adult-like system in English with exposure and practice.

Phoneme accuracy as a function of mode of communication in pediatric cochlear implantation

This study investigated phoneme accuracy in cochlear implanted (CI) children using auditory-oral versus total communication modes of communication as a function of sentences containing words easily guessed by the linguistic environment versus words unaided by the linguistic environment. A total of 173 children from 33 different states and five Canadian provinces participated. Communication mode was assessed through a parental rating scale inquiring how emphasis was placed on speech and auditory development in the educational setting. Thirty-six sentences varying in length (three, five, or seven syllables) contained a key word predicting speech intelligibility in deaf children. Average phoneme accuracy was 76.38% for the CI children. Across CI participants, vowel productions were more accurate (81.30%) than consonants (73.64%). Overall, phoneme accuracy was higher for key words in sentences containing more contextual information (78.57%) than lower contextual information (74.21%). Consonants were more accu...

Time-domain isolated phoneme classification using reconstructed phase spaces

This paper introduces a novel time-domain approach to modeling and classifying speech phoneme waveforms. The approach is based on statistical models of reconstructed phase spaces, which offer significant theoretical benefits as representations that are known to be topologically equivalent to the state dynamics of the underlying production system. The lag and dimension parameters of the reconstruction process for speech are examined in detail, comparing common estimation heuristics for these parameters with corresponding maximum likelihood recognition accuracy over the TIMIT data set. Overall accuracies are compared with a Mel-frequency cepstral baseline system across five different phonetic classes within TIMIT, and a composite classifier using both cepstral and phase space features is developed. Results indicate that although the accuracy of the phase space approach by itself is still currently below that of baseline cepstral methods, a combined approach is capable of increasing speaker independent phoneme accuracy.

Comparison of psychophysical and T-NRT levels in children using N24 Contour™ vs. children using N24 M or K. Phonemic development

Objective To compare intraoperative T-NRT (neural response telemetry) using perimodiolar Nucleus 24 Contour™ Cochlear Implant (C.I.) with the ones obtained using straight Nucleus 24 M or K. To compare the levels T (threshold) and C (comfort) in the psychofisic measurements made at the first fitting and 3 months later. To determine differences in phonological development. Study design Retrospective, observational, descriptive. Statistical analysis: Student t-test. A sample of 22 children was taken: Intrasurgical T-NRT was obtained by visual technique. The stimulus electrodes were 15, 10 and 5. T and C levels were obtained using the SPEAK strategy. Analysis of phoneme accuracy was obtained at 12 and 24 months of C.I. use in children under 6 years old. Results For the perimodiolar group T-NRT was obtained between 6 and 13 levels below the obtained for the group with straight electrode. For the selected electrodes, T and C level at the first fitting was obtained between 10 and 18 levels below and 3 months later the C levels remain 10 units below the one obtained with straight electrode. These group had higher growth rate of phonemes recognition at 24 months of C.I. use. Conclusions Intrasurgical T-NRT was obtained with smaller level of stimulation using perimodiolar electrode than using straight electrode ( p<0.05) The T and C level at the first fitting and after 3 months later were smaller for the perimodiolar group in relation to the straight group ( p<0.05). The dynamic range was similar in both groups. The development of higher growth rate of phoneme recognition was found in young children using perimodiolar array at 24 months of C.I. use; being necessary to carry out this analysis in a bigger population.

Assessing text-to-phoneme mapping strategies in speaker independent isolated word recognition

A phonetic transcription of the vocabulary, i.e., a lexicon, is needed in sub-word based speech recognition and text-to-speech systems. Decision trees and neural networks have successfully been used for creating lexicons on-line from an open vocabulary. We briefly review these methods and compare them in detail in the text-to-phoneme mapping task as part of a phoneme based speaker independent speech recognizer. The decision tree and neural network based methods were first evaluated in terms of phoneme accuracy and then in extensive speech recognition tests. American english dictionaries and speech databases were used in all experiments. The decision tree based method achieved high phoneme accuracies when the training material covered the test vocabulary well. In typical speech recognition tests, the recognition rates obtained using the decision tree based lexicons were close to the baseline that was obtained using accurate transcriptions. Although the lexicons obtained using neural networks resulted in somewhat lower baseline recognition rates, they provided slightly better results in generalization tests. Moreover, when the neural network based mappings were appended with a look-up table comprising the most likely vocabulary items, which would be the practical set-up, their performance increased significantly. The main advantage of neural networks over decision trees is their low memory consumption.

Analysis of the Phonological Skills of Children with Down Syndrome from Single Word and Connected Speech Samples

Phonological analysis of the speech of five children with Down syndrome was compared across single word samples elicited from tests and connected speech samples elicited during play. Analyses indicated that connected speech samples provided fewer total words and word tokens and, in some cases, failed to target certain later developing phonemes. CVCs were similarly represented across single word and connected speech samples, but differences across the sampling conditions included an over‐representation of clusters and an under‐representation of VC and CV forms in single word samples. Overall accuracy for vowels was greater than for consonant production, but no difference in phoneme accuracy was found across the conditions. Differences in phonetic inventories for some subjects indicated the possibility that children may, in their connected speech, avoid phonemes which they have yet to master. Phonological process analysis indicated that differences across the two conditions were restricted to their relative impact on each subject's overall intelligibility. Both similarities and differences with previous studies comparing sampling conditions with other populations are discussed, as are clinical implications.