Evaluating the consistency of lenition measures: Neural networks' posterior probability, intensity velocity, and duration.

Predictions of gradient degree of lenition of voiceless and voiced stops in a corpus of Argentine Spanish are evaluated using three acoustic measures (minimum and maximum intensity velocity and duration) and two recurrent neural network (Phonet) measures (posterior probabilities of sonorant and continuant phonological features). While mixed and inconsistent predictions were obtained across the acoustic metrics, sonorant and continuant probability values were consistently in the direction predicted by known factors of a stop's lenition with respect to its voicing, place of articulation, and surrounding contexts. The results suggest the effectiveness of Phonet as an additional or alternative method of lenition measurement. Furthermore, this study has enhanced the accessibility of Phonet by releasing the trained Spanish Phonet model used in this study and a pipeline with step-by-step instructions for training and inferencing new models.

Disentangling Effects of Memory Storage and Inter-articulator Coordination on Generalization in Speech Motor Sequence Learning.

Generalization in motor control is the extent to which motor learning affects movements in situations different than those in which it originally occurred. Recent data on orofacial speech movements indicates that motor sequence learning generalizes to novel syllable sequences containing phonotactically illegal, but previously practiced, consonant clusters. Practicing an entire syllable, however, results in even larger performance gains compared to practicing just its clusters. These patterns of generalization could reflect language-general changes in phonological memory storage and/or inter-articulator coordination during motor sequence learning. To disentangle these factors, we conducted two experiments in which talkers intensively practiced producing novel syllables containing illegal onset and coda clusters over two consecutive days. During the practice phases of both experiments, we observed that, through repetition, talkers gradually produced the syllables with fewer errors, indicative of learning. After learning, talkers were tested for generalization to single syllables (Experiment 1) or syllable pairs (Experiment 2) that overlapped to varying degrees with the practiced syllables. Across both experiments, we found that performance improvements from practicing syllables with illegal clusters partially generalized to novel syllables that contained those clusters, but performance was more error prone if the clusters occurred in a different syllable position (onset versus coda) as in practice, demonstrating that inter-articulator coordination is contextually sensitive. Furthermore, changing the position of a cluster was found to be more deleterious to motor performance during the production of the second syllables in syllable pairs, which required talkers to store more phonological material in memory prior to articulation, compared to single syllables. This interaction effect reveals a complex interplay between memory storage and inter-articulator coordination on generalization in speech motor sequence learning.

From sonority hierarchy to posterior probability as a measure of lenition: The case of Spanish stops.

A deep learning Phonet model was evaluated as a method to measure lenition. Unlike quantitative acoustic methods, recurrent networks were trained to recognize the posterior probabilities of sonorant and continuant phonological features in a corpus of Argentinian Spanish. When applied to intervocalic and post-nasal voiced and voiceless stops, the approach yielded lenition patterns similar to those previously reported. Further, additional patterns also emerged. The results suggest the validity of the approach as an alternative or addition to quantitative acoustic measures of lenition.

An investigation of interference between electromagnetic articulography and electroglottography.

The present study tested whether there is cross-interference between electromagnetic articulography (EMA) and electroglottography (EGG) during the acquisition of kinematic speech data. In experiments 1A and 1B, EMA sensors were calibrated with and without EGG electrodes present in the EMA field. In experiment 2, EMA was used to record lip, tongue, and jaw movements for one male speaker and one female speaker, with and without simultaneous EGG recording. Collectively, the results provide no evidence of signal artifacts in either direction, suggesting that EMA and EGG technology can be combined to reliably assess laryngeal and supralaryngeal motor coordination in speech.

Categorical Perception of Mandarin Pitch Directions by Cantonese-Speaking Musicians and Non-musicians.

Purpose: This study is to investigate whether Cantonese-speaking musicians may show stronger CP than Cantonese-speaking non-musicians in perceiving pitch directions generated based on Mandarin tones. It also aims to examine whether musicians may be more effective in processing stimuli and more sensitive to subtle differences caused by vowel quality. Methods: Cantonese-speaking musicians and non-musicians performed a categorical identification and a discrimination task on rising and falling continua of fundamental frequency generated based on Mandarin level, rising and falling tones on two vowels with nine duration values. Results: Cantonese-speaking musicians exhibited a stronger categorical perception (CP) of pitch contours than non-musicians based on the identification and discrimination tasks. Compared to non-musicians, musicians were also more sensitive to the change of stimulus duration and to the intrinsic F 0 in pitch perception in pitch processing. Conclusion: The CP was strengthened due to musical experience and musicians benefited more from increased stimulus duration and were more efficient in pitch processing. Musicians might be able to better use the extra time to form an auditory representation with more acoustic details. Even with more efficiency in pitch processing, musicians' ability to detect subtle pitch changes caused by intrinsic F 0 was not undermined, which is likely due to their superior ability to process temporal information. These results thus suggest musicians may have a great advantage in learning tones of a second language.

How musical experience affects tone perception efficiency by musicians of tonal and non-tonal speakers?

PURPOSE: To investigate if, regardless of language background (tonal or non-tonal), musicians may show stronger CP than non-musicians; To examine if native speakers of English (English or non-tonal musicians henceforth) or Mandarin Chinese (Mandarin or tonal musicians henceforth) can better accommodate multiple functions of the same acoustic cue and if musicians' sensitivity to pitch of lexical tones comes at the cost of slower processing. METHOD: English and Mandarin Musicians and non-musicians performed a categorical identification and a discrimination task on rising and falling continua of fundamental frequency on two vowels with 9 duration values. RESULTS: Non-tonal musicians exhibited significantly stronger categorical perception of pitch contour than non-tonal non-musicians. However, tonal musicians did not consistently perceive the two types of pitch directions more categorically than tonal non-musicians. Both tonal and non-tonal musicians also benefited more from increasing stimulus duration in processing pitch changes than non-musicians and they generally require less time for pitch processing. Musicians were also more sensitive to intrinsic F0 in pitch perception and differences of pitch types. CONCLUSION: The effect of musical training strengthens categorical perception more consistently in non-tonal speakers than tonal speakers. Overall, musicians benefit more from increased stimulus duration, due perhaps to their greater sensitivity to temporal information, thus allowing them to be better at forming a more robust auditory representation and matching sounds to internalized memory templates. Musicians also attended more to acoustic details such as intrinsic F0 and pitch types in pitch processing, and yet, overall, their categorization of pitch was not compromised by traces of these acoustic details from their auditory short-term working memory. These findings may lead to a better understanding of pitch perception deficits in special populations, particularly among individuals diagnosed with autism spectrum disorder (ASD).

Visual analog of the acoustic amplitude envelope benefits speech perception in noise.

The nature of the visual input that integrates with the audio signal to yield speech processing advantages remains controversial. This study tests the hypothesis that the information extracted for audiovisual integration includes co-occurring suprasegmental dynamic changes in the acoustic and visual signal. English sentences embedded in multi-talker babble noise were presented to native English listeners in audio-only and audiovisual modalities. A significant intelligibility enhancement with the visual analogs congruent to the acoustic amplitude envelopes was observed. These results suggest that dynamic visual modulation provides speech rhythmic information that can be integrated online with the audio signal to enhance speech intelligibility.

Effects of stimulus duration and vowel quality in cross-linguistic categorical perception of pitch directions.

We investigated categorical perception of rising and falling pitch contours by tonal and non-tonal listeners. Specifically, we determined minimum durations needed to perceive both contours and compared to those of production, how stimuli duration affects their perception, whether there is an intrinsic F0 effect, and how first language background, duration, directions of pitch and vowel quality interact with each other. Continua of fundamental frequency on different vowels with 9 duration values were created for identification and discrimination tasks. Less time is generally needed to effectively perceive a pitch direction than to produce it. Overall, tonal listeners' perception is more categorical than non-tonal listeners. Stimuli duration plays a critical role for both groups, but tonal listeners showed a stronger duration effect, and may benefit more from the extra time in longer stimuli for context-coding, consistent with the multistore model of categorical perception. Within a certain range of semitones, tonal listeners also required shorter stimulus duration to perceive pitch direction changes than non-tonal listeners. Finally, vowel quality plays a limited role and only interacts with duration in perceiving falling pitch directions. These findings further our understanding on models of categorical perception, the relationship between speech perception and production, and the interaction between the perception of tones and vowel quality.

Effects of production training and perception training on lexical tone perception--A behavioral and ERP study.

The present study recorded both behavioral data and event-related brain potentials to examine the effectiveness of a perception-only training and a perception-plus-production training procedure on the intentional and unintentional perception of lexical tone by native English listeners. In the behavioral task, both the perception-only and the perception-plus-production groups improved on the tone discrimination abilities after the training session. Moreover, the participants in both groups generalized the improvements gained through the trained stimuli to the untrained stimuli. In the ERP task, the Mismatch Negativity was smaller in the post-training task than in the pre-training task. However, the two training groups did not differ in tone processing at the intentional or unintentional level after training. These results suggest that the employment of the motor system does not specifically benefit the tone perceptual skills. Furthermore, the present study investigated whether some tone pairs are more easily confused than others by native English listeners, and whether the order of tone presentation influences non-native tone discrimination. In the behavioral task, Tone2-Tone1 (rising-level) and Tone2-Tone4 (rising-falling) were the most difficult tone pairs, while Tone1-Tone2 and Tone4-Tone2 were the easiest tone pairs, even though they involved the same tone contrasts respectively. In the ERP task, the native English listeners had good discrimination when Tone2 and Tone4 were embedded in strings of Tone1, while poor discrimination when Tone1 was inserted in the context of Tone2 or Tone4. These asymmetries in tone perception might be attributed to the interference of native intonation system and can be altered by training.

Changes in oscillatory brain networks after lexical tone training.

Learning foreign speech contrasts involves creating new representations of sound categories in memory. This formation of new memory representations is likely to involve changes in neural networks as reflected by oscillatory brain activity. To explore this, we conducted time-frequency analyses of electro-encephalography (EEG) data recorded in a passive auditory oddball paradigm using Thai language tones. We compared native speakers of English (a non-tone language) and native speakers of Mandarin Chinese (a tone language), before and after a two-day laboratory training. Native English speakers showed a larger gamma-band power and stronger alpha-band synchrony across EEG channels than the native Chinese speakers, especially after training. This is compatible with the view that forming new speech categories on the basis of unfamiliar perceptual dimensions involves stronger gamma activity and more coherent activity in alpha-band networks than forming new categories on the basis of familiar dimensions.

Speech perception among school-aged skilled and less skilled readers.

This study investigated the relationship between speech perception, phonological processing and reading skills among school-aged children classified as 'skilled' and 'less skilled' readers based on their ability to read words, decode non-words, and comprehend short passages. Three speech perception tasks involving categorization of speech continua differing in voicing, place and manner of articulation were administered and compared to phonological processing skills in phonological awareness, speeded naming and verbal short-term memory. The results obtained suggested that (a) speech categorization among skilled readers differed from that of less skilled readers, (b) speech perception skills were associated with both reading and phonological processing skills among both skilled and less skilled readers, however, (c) a strong association between speeded naming and both word and passage reading skills found among skilled readers was absent among less skilled readers. These results suggested that phonological representations and/or activation may not be as well developed in less skilled readers.

Thai lexical tone perception in native speakers of Thai, English and Mandarin Chinese: an event-related potentials training study.

BACKGROUND: Tone languages such as Thai and Mandarin Chinese use differences in fundamental frequency (F0, pitch) to distinguish lexical meaning. Previous behavioral studies have shown that native speakers of a non-tone language have difficulty discriminating among tone contrasts and are sensitive to different F0 dimensions than speakers of a tone language. The aim of the present ERP study was to investigate the effect of language background and training on the non-attentive processing of lexical tones. EEG was recorded from 12 adult native speakers of Mandarin Chinese, 12 native speakers of American English, and 11 Thai speakers while they were watching a movie and were presented with multiple tokens of low-falling, mid-level and high-rising Thai lexical tones. High-rising or low-falling tokens were presented as deviants among mid-level standard tokens, and vice versa. EEG data and data from a behavioral discrimination task were collected before and after a two-day perceptual categorization training task. RESULTS: Behavioral discrimination improved after training in both the Chinese and the English groups. Low-falling tone deviants versus standards elicited a mismatch negativity (MMN) in all language groups. Before, but not after training, the English speakers showed a larger MMN compared to the Chinese, even though English speakers performed worst in the behavioral tasks. The MMN was followed by a late negativity, which became smaller with improved discrimination. The High-rising deviants versus standards elicited a late negativity, which was left-lateralized only in the English and Chinese groups. CONCLUSION: Results showed that native speakers of English, Chinese and Thai recruited largely similar mechanisms when non-attentively processing Thai lexical tones. However, native Thai speakers differed from the Chinese and English speakers with respect to the processing of late F0 contour differences (high-rising versus mid-level tones). In addition, native speakers of a non-tone language (English) were initially more sensitive to F0 onset differences (low-falling versus mid-level contrast), which was suppressed as a result of training. This result converges with results from previous behavioral studies and supports the view that attentive as well as non-attentive processing of F0 contrasts is affected by language background, but is malleable even in adult learners.

Effects of native language and training on lexical tone perception: an event-related potential study.

Tone languages such as Thai use pitch differences to distinguish lexical meaning. Previous behavioral studies have reported that naïve listeners can discriminate among lexical tones, but that native language background affects performance. The present study uses ERPs to determine whether native speakers of a tone language (Mandarin Chinese) and of a non-tone language (English) differ in their pre-attentive discrimination among Thai lexical tones, and whether training has a different effect in these two language groups. EEGs were obtained from 10 native Mandarin Chinese speakers, 10 English and 10 Thai speakers in an oddball paradigm: The Thai syllable [k(h)a:] pronounced with a high rising or low falling tone, was presented as an infrequent deviant amidst a standard mid level tone [k(h)a:] syllable, while participants watched a silent movie. Next, the Chinese and English participants completed a 2-day perceptual identification training on the mid level and low falling tones, and returned for a post training EEG. The low falling tone deviant elicited a Mismatch Negativity (MMN) in all participant groups before and after training; the high rising deviant elicited no, or a smaller, MMN, which became larger after training only in the English group. The high rising deviant also elicited a later negativity (350-650 ms) versus the mid level standard, which decreased after training in the Chinese group. These results suggest that non-Thai speakers can pre-attentively discriminate among Thai tones, but are sensitive to different physical properties of the tones, depending on their native language. English speakers are more sensitive to early pitch differences, whereas native speakers of Mandarin Chinese are more sensitive to the (later) pitch contour.

Native Thai speakers' acquisition of English word stress patterns.

The influence of syllabic structure, lexical class and stress patterns of known words on the acquisition of the English stress system was investigated in ten native Thai speakers. All participants were adult learners of English with an average length of residence in the US of 1.4 years. They were asked to produce and give perceptual judgments on 40 English non-words of varying syllabic structures in noun and verb sentence frames. Results of the production data suggested that syllables with a long vowel attracted stress more often than syllables containing a short vowel and nouns received initial stress more often than verbs. Additionally, regression analyses with the three factors as predictors suggested that Thai participants' pattern of stress assignment on non-words was significantly influenced by the stress patterns of phonologically similar real words. These results were compared and contrasted to those found in previous work with Spanish-English and Korean-English bilinguals.

Factors affecting stress placement for English nonwords include syllabic structure, lexical class, and stress patterns of phonologically similar words.

Seventeen native English speakers participated in an investigation of language users' knowledge of English main stress patterns. First, they produced 40 two-syllable nonwords of varying syllabic structure as nouns and verbs. Second, they indicated their preference for first or second syllable stress of the same words in a perception task. Finally, they indicated words they considered to be phonologically similar to the nonwords. Analyses of variance on the production and perception data indicated that both syllabic structure and lexical class (noun or verb) had an effect on main stress assignment. In logistic regression analyses on the production and perception responses. predictions of stress placement made by (1) syllable structure, (2) lexical class, and (3) stress patterns of phonologically similar words all contributed significantly and uniquely to the prediction of main stress assignment. The results indicate that phonological theories of English word stress need to allow for multiple, competing, probabilistic factors in accounts of main stress placement including syllabic structure (most notably vowel length), lexical class, and stress patterns of phonologically similar words.