Articulatory Features Research Articles

350 Investigating the Architecture of Speech Processing Pathways in the Brain

OBJECTIVES/GOALS: Speech production requires mapping between sound-based and motor-based neural representations of a word – accomplished by learning internal models. However, the neural bases of these internal models remain unclear. The aim of this study is to provide experimental evidence for these internal models in the brain during speech production. METHODS/STUDY POPULATION: 16 healthy human adults were recruited for this electrooculography speech study. 20 English pseudowords were designed to vary on confusability along specific features of articulation (place vs manner). All words were controlled for length and voicing. Three task conditions were performed: speech perception, covert and overt speech production. EEG was recorded using a 64-channel Biosemi ActiveTwo system. EMG was recorded on the orbicularis orbis inferior and neck strap muscles. Overt productions were recorded with a high-quality microphone to determine overt production onset. EMG during was used to determine covert production onset. Neuroimaging: Representational Similarity Analysis (RSA), was used to probe the sound- and motor-based neural representations over sensors and time for each task. RESULTS/ANTICIPATED RESULTS: Production (motor) and perception (sound) neural representations were calculated using a cross-validated squared Euclidean distance metric. The RSA results in the speech perception task show a strong selectivity around 150ms, which is compatible with recent human electrocorticography findings in human superior temporal gyrus. Parietal sensors showed a large difference for motor-based neural representations, indicating a strong encoding for production related processes, as hypothesized by previous studies on the ventral and dorsal stream model of language. Temporal sensors, however, showed a large change for both motor- and sound-based neural representations. This is a surprising result since temporal regions are believed to be primarily engaged in perception (sound-based) processes. DISCUSSION/SIGNIFICANCE: This study used neuroimaging (EEG) and advanced multivariate pattern analysis (RSA) to test models of production (motor-) and perception (sound-) based neural representations in three different speech task conditions. These results show strong feasibility of this approach to map how the perception and production processes interact in the brain.

Characteristics of articulatory gestures in stuttered speech: A case study using real-time magnetic resonance imaging

IntroductionMost of the previous articulatory studies of stuttering have focussed on the fluent speech of people who stutter. However, to better understand what causes the actual moments of stuttering, it is necessary to probe articulatory behaviors during stuttered speech. We examined the supralaryngeal articulatory characteristics of stuttered speech using real-time structural magnetic resonance imaging (RT-MRI). We investigated how articulatory gestures differ across stuttered and fluent speech of the same speaker. MethodsVocal tract movements of an adult man who stutters during a pseudoword reading task were recorded using RT-MRI. Four regions of interest (ROIs) were defined on RT-MRI image sequences around the lips, tongue tip, tongue body, and velum. The variation of pixel intensity in each ROI over time provided an estimate of the movement of these four articulators. ResultsAll disfluencies occurred on syllable-initial consonants. Three articulatory patterns were identified. Pattern 1 showed smooth gestural formation and release like fluent speech. Patterns 2 and 3 showed delayed release of gestures due to articulator fixation or oscillation respectively. Block and prolongation corresponded to either pattern 1 or 2. Repetition corresponded to pattern 3 or a mix of patterns. Gestures for disfluent consonants typically exhibited a greater constriction than fluent gestures, which was rarely corrected during disfluencies. Gestures for the upcoming vowel were initiated and executed during these consonant disfluencies, achieving a tongue body position similar to the fluent counterpart. ConclusionDifferent perceptual types of disfluencies did not necessarily result from distinct articulatory patterns, highlighting the importance of collecting articulatory data of stuttering. Disfluencies on syllable-initial consonants were related to the delayed release and the overshoot of consonant gestures, rather than the delayed initiation of vowel gestures. This suggests that stuttering does not arise from problems with planning the vowel gestures, but rather with releasing the overly constricted consonant gestures.

Classification of accurate and misarticulated /ɑr/ for ultrasound biofeedback using tongue part displacement trajectories

ABSTRACT Ultrasound biofeedback therapy (UBT), which incorporates real-time imaging of tongue articulation, has demonstrated generally positive speech remediation outcomes for individuals with residual speech sound disorder (RSSD). However, UBT requires high attentional demands and may therefore benefit from a simplified display of articulation targets that are easily interpretable and can be compared to real-time articulation. Identifying such targets requires automatic quantification and analysis of movement features relevant to accurate speech production. Our image-analysis program TonguePART automatically quantifies tongue movement as tongue part displacement trajectories from midsagittal ultrasound videos of the tongue, with real-time capability. The present study uses such displacement trajectories to compare accurate and misarticulated American-English rhotic /ɑr/ productions from 40 children, with degree of accuracy determined by auditory perceptual ratings. To identify relevant features of accurate articulation, support vector machine (SVM) classifiers were trained and evaluated on several candidate data representations. Classification accuracy was up to 85%, indicating that quantification of tongue part displacement trajectories captured tongue articulation characteristics that distinguish accurate from misarticulated production of /ɑr/. Regression models for perceptual ratings were also compared. The simplest data representation that retained high predictive ability, demonstrated by high classification accuracy and strong correlation between observed and predicted ratings, was displacements at the midpoint of /r/ relative to /ɑ/ for the tongue dorsum and blade. This indicates that movements of the dorsum and blade are especially relevant to accurate production of /r/, suggesting that a predictive parameter and biofeedback target based on this data representation may be usable for simplified UBT.

Speech sound discrimination by Mongolian gerbils

The present study establishes the Mongolian gerbil (Meriones unguiculatus) as a model for investigating the perception of human speech sounds. We report data on the discrimination of logatomes (CVCs - consonant-vowel-consonant combinations with outer consonants /b/, /d/, /s/ and /t/ and central vowels /a/, /aː/, /ɛ/, /eː/, /ɪ/, /iː/, /ɔ/, /oː/, /ʊ/ and /uː/, VCVs - vowel-consonant-vowel combinations with outer vowels /a/, /ɪ/ and /ʊ/ and central consonants /b/, /d/, /f/, /g/, /k/, /l/, /m/, /n/, /p/, /s/, /t/ and /v/) by gerbils. Four gerbils were trained to perform an oddball target detection paradigm in which they were required to discriminate a deviant CVC or VCV in a sequence of CVC or VCV standards, respectively. The experiments were performed with an ICRA-1 noise masker with speech-like spectral properties, and logatomes of multiple speakers were presented at various signal-to-noise ratios. Response latencies were measured to generate perceptual maps employing multidimensional scaling, which visualize the gerbils’ internal maps of the sounds. The dimensions of the perceptual maps were correlated to multiple phonetic features of the speech sounds for evaluating which features of vowels and consonants are most important for the discrimination. The perceptual representation of vowels and consonants in gerbils was similar to that of humans, although gerbils needed higher signal-to-noise ratios for the discrimination of speech sounds than humans. The gerbils’ discrimination of vowels depended on differences in the frequencies of the first and second formant determined by tongue height and position. Consonants were discriminated based on differences in combinations of their articulatory features. The similarities in the perception of logatomes by gerbils and humans renders the gerbil a suitable model for human speech sound discrimination.

Acoustic and Articulatory Characteristics of English Semivowels /ɹ, l, w/ Produced by Adult Second-Language Speakers.

This study presents the results of acoustic and kinematic analyses of word-initial semivowels (/ɹ, l, w/) produced by second-language (L2) speakers of English whose native language is Korean. In addition, the relationship of acoustic and kinematic measures to the ratings of foreign accent was examined by correlation analyses. Eleven L2 speakers and 10 native speakers (first language [L1]) of English read The Caterpillar passage. Acoustic and kinematic data were simultaneously recorded using an electromagnetic articulography system. In addition to speaking rate, two acoustic measures (ratio of third-formant [F3] frequency to second-formant [F2] frequency and duration of steady states of F2) and two kinematic measures (lip aperture and duration of lingual maximum hold) were obtained from individual target sounds. To examine the degree of contrast among the three sounds, acoustic and kinematic Euclidean distances were computed on the F2-F3 and x-y planes, respectively. Compared with L1 speakers, L2 speakers exhibited a significantly slower speaking rate. For the three semivowels, L2 speakers showed a reduced F3/F2 ratio during constriction, increased lip aperture, and reduced acoustic Euclidean distances among semivowels. Additionally, perceptual ratings of foreign accent were significantly correlated with three measures: duration of steady F2, acoustic Euclidean distance, and kinematic Euclidean distance. The findings provide acoustic and kinematic evidence for challenges that L2 speakers experience in the production of English semivowels, especially /ɹ/ and /w/. The robust and consistent finding of reduced contrasts among semivowels and their correlations with perceptual accent ratings suggests using sound contrasts as a potentially effective approach to accent modification paradigms.

Сравнительно-сопоставительная характеристика акустических особенностей звука [а] в татарском и казахском языках

The article is devoted to the experimental study of the articulation characteristics of the sound systems of the related Tatar and Kazakh languages using the example of the vowel sound [a]. The aim of the study is to identify the specifics of using the studied sound in different positions from the point of view of its articulatory characteristics. Despite the fact that a lot of works have been devoted to studying the phonetic systems of the compared languages, the issue of comparative analysis of the length, strength and frequency of the studied sound by the material of the Tatar and Kazakh languages using the Speech Analyzer program is tackled for the first time, which is the scientific originality of this research. As a result of the linguistic material analysis, it was possible to determine the specifics of the indicators of the length, strength and frequency of the sound [a] in different positions by the material of the Tatar and Kazakh languages.

Discovering Lexical Similarity Using Articulatory Feature-Based Phonetic Edit Distance

Lexical Similarity (LS) between two languages uncovers many interesting linguistic insights such as genetic relationship, mutual intelligibility, and the usage of one's vocabulary into other. There are various methods through which LS is evaluated. In the same regard, this paper presents a method of Phonetic Edit Distance (PED) that uses a soft comparison of letters using the articulatory features associated with them. The system converts the words into the corresponding International Phonetic Alphabet (IPA), followed by the conversion of IPA into its set of articulatory features. Later, the lists of the set of articulatory features are compared using the proposed method. As an example, PED gives edit distance of German word vater and Persian word pidar as 0.82; and similarly, Hebrew word shalom and Arabic word salaam as 0.93, whereas for a juxtapose comparison, their IPA based edit distances are 4 and 2 respectively. Experiments are performed with six languages (Arabic, Hindi, Marathi, Persian, Sanskrit, and Urdu). In this regard, we extracted part of speech wise word-lists from the Universal Dependency corpora and evaluated the LS for every pair of language. Thus, with the proposed approach, we find the genetic affinity, similarity, and borrowing/loan-words despite having script differences and sound variation phenomena among these languages.

Anticipation, Simultaneity/Consecutivity and Distinctive Features in Phonology

The notion of distinctive features has had a firm position in phonology since the time of the Prague Linguistic Circle and especially that of one of its representatives, Roman Jakobson, whose well-known delimitation of a phoneme as “a bundle of distinctive features” (Jakobson, 1962, p. 421), that is, a set of simultaneous distinctive features, has inspired many scholars. Jakobson’s attempt “to analyse the distribution of distinctive features along two axes: that of simultaneity and that of successiveness” (ibid., p. 435) helped cover several phonetic and/or phonological processes and phenomena. Distinctive features, although theoretical constructs (Giegerich, 1992, p. 89), reflect phonetic, that is, articulatory and acoustic, properties of sounds. In the flow of speech, some features tend to influence the neighbouring phonemes. Sometimes speech organs produce something that the brain just ‘plans’ to produce (anticipatory speech errors). There are situations where it seems as if the successive organization of phonemes went hand in hand with the simultaneous nature of certain articulatory characteristics of those phonemes (the transgression of consonants and inherence of vowels in Romportl’s theory), or the given feature seems to be anticipated by the preceding segment. This is the case with nasalization and/or anticipatory coarticulation, as well as regressive (anticipatory) assimilation. In addition, simultaneity/consecutivity is a decisive criterion for the difference between the so-called complex segments, as specified in Feature Geometry, and simple segments (Duanmu, 2009). Moreover, the phonological opposition of simultaneity- successivity (that is, consecutivity) itself functions as a feature making a difference between segmental and suprasegmental elements in the sound system of a language, as was first mentioned by Harris (1944), later indicated by Jakobson (1962) and then fully developed by Sabol (2007, 2012).

Domain-Adversarial Based Model with Phonological Knowledge for Cross-Lingual Speech Recognition

Phonological-based features (articulatory features, AFs) describe the movements of the vocal organ which are shared across languages. This paper investigates a domain-adversarial neural network (DANN) to extract reliable AFs, and different multi-stream techniques are used for cross-lingual speech recognition. First, a novel universal phonological attributes definition is proposed for Mandarin, English, German and French. Then a DANN-based AFs detector is trained using source languages (English, German and French). When doing the cross-lingual speech recognition, the AFs detectors are used to transfer the phonological knowledge from source languages (English, German and French) to the target language (Mandarin). Two multi-stream approaches are introduced to fuse the acoustic features and cross-lingual AFs. In addition, the monolingual AFs system (i.e., the AFs are directly extracted from the target language) is also investigated. Experiments show that the performance of the AFs detector can be improved by using convolutional neural networks (CNN) with a domain-adversarial learning method. The multi-head attention (MHA) based multi-stream can reach the best performance compared to the baseline, cross-lingual adaptation approach, and other approaches. More specifically, the MHA-mode with cross-lingual AFs yields significant improvements over monolingual AFs with the restriction of training data size and, which can be easily extended to other low-resource languages.

Who is Right? A Word-Identification-in-Noise Test for Young Children Using Minimal Pair Distracters.

Many children have difficulties understanding speech. At present, there are few assessments that test for subtle impairments in speech perception with normative data from U.K. children. We present a new test that evaluates children's ability to identify target words in background noise by choosing between minimal pair alternatives that differ by a single articulatory phonetic feature. This task (a) is tailored to testing young children, but also readily applicable to adults; (b) has minimal memory demands; (c) adapts to the child's ability; and (d) does not require reading or verbal output. We tested 155 children and young adults aged from 5 to 25 years on this new test of single word perception. Speech-in-noise abilities in this particular task develop rapidly through childhood until they reach maturity at around 9 years of age. We make this test freely available and provide associated normative data. We hope that it will be useful to researchers and clinicians in the assessment of speech perception abilities in children who are hard of hearing or have developmental language disorder, dyslexia, or auditory processing disorder. https://doi.org/10.23641/asha.17155934.

Yugan Khanty Phonetic Features Based on the Experimental Data

The paper deals with the phonetics of Yugan idiom of Surgut Khanty. The research is a part of the project aimed at describing Surgut Khanty phonetics. The Yugan idiom has significant differences from the Tromyegan idiom described before.
 The analysis is based on the data collected during the expedition to the settlement Ugut in 2019. The experimental part includes 130 words list read out three times by four native speakers from different traditional settlements on the Malyi Yugan river and on the Bolshoi Yugan river.
 The research was conducted using experimental techniques: Praat, Emu-SDMS software. The main technique was a formant analysis that deals with resonant frequencies in vowel spectra to obtain data on articulation features. Statistical evaluations and visualization were established via R programming language.
 We found differences between the Malyi Yugan river and the Bol’shoi Yugan river idioms. 12 vowel phonemes were found in the Malyi Yugan idiom. Compared to the Tromyegan system the phoneme /ɔ/ (traditionally /ȯ̆/) is absent. It was replaced by /ɛ/ (traditionally /ȧ̆/) or /o/ (traditionally /ŏ/). The phoneme u̇ described in previous literature on the topic disappeared and was replaced by /iː/.
 The Bolshoi Yugan vowel system includes these phonemes and also diphthongs [ui], [ɔɛ]. They appear after [k] where etimological u̇, ȯ̆ used to be. They probably are the realizations of the phonemes /iː/, /ɛ/ in the position after labialized k, which has become a phoneme.
 Non-initial [w] is reported to be specific Jugan feature and appears to have parallels in Tromyegan idiom too. It is an evidence for the rearranging of the Surgut idioms. In this pronunciation type /w/ is realized as a labial approximant in an initial position and after not-rounded vowels in a non-final position. After not-rounded vowels in a final position it comes as an initial-voiced fricative evoking preceding vowel diphthongization. After rounded vowels it is labiovelar [γʷ] or non-syllabic [ʊ] (before consonants). This pronunciation type is similar to the Tromyegan type, but it differs from the Pim type where /w/ comes as a labial approximant consistently.
 The disappearance of labial fricatives is a new phenomenon which has not been described properly. Territorial and social factors for this process are given. The Malyi Yugan speakers use lateral fricatives /ł/, /ʎ̥/ and the Bolshoi Yugan speakers replace it by /t/, /c͡c̦/. In the settlement Ugut where Bolshoi and Malyi Yugan natives contact in Russian-spoken environment both variants are used with t-pronunciation evaluated by speakers as new and declining from the ‘'right’ speech.

The Poet as Cyber-Linguistic Programmer of Her Age: McCaffery

This article posits that Steve McCaffery's radical poetico-graphic experimentation, from the 1970s through to the present day, is an exemplary site at which innovations in poetics and technology converge. At stake in McCaffery's multifaceted achievement are, among other issues: the composition of the poetic page as a screen, with its integral design and integrity; the transition between analog and digital print organisations, in thinking as in technology; the history of printing, as evidenced by his deployment of the transitional technology of the IBM Selectric typewriter in the composition of images as well as texts; and the notion that graphics (figures, cartoons, maps, and charts – with and without words) can be imbued with the figuration and articulatory nuance characteristic of verse. Throughout McCaffery's double compilation volume, Seven Pages Missing (2000 and 2002), the poet explicitly registers the transformations in communications, information, and technology that he could not help but experience around him. His poetic experiments are concurrent with luminous introductions to the cybernetic universe furnished by the likes of Anthony Wilden and Douglas Hofstadter. McCaffery's experiments, in their boldness and scope, also shed light on the venerable, ongoing cultural play between analog and digital: not period-specific or bound to particular technologies of calculation, communications, and reproduction.

Religious Arabic exclamations in Persian talks-in-interaction: a micro-analytic approach to contexts and senses

ABSTRACT Through centuries many Arabic expressions have been integrated into Persian. Islamic Arabic exclamations also made their way into everyday Persian and social life of Iranians, to the extent that today few might even note their origin and distinction. The current study was purposed to investigate the illocutionary meanings religious exclamations have acquired in the modern society of Iranian Shiites. It also aimed at examining how different functions are served by the same exclamations but with various articulatory suprasegmental features. Adopting an inductively micro-analytic qualitative approach to the study of talks-in-interaction, naturally occurring conversations in vernacular Persian were observed and recorded. A sample of thirty actual conversations were systematically transcribed and analysed using Jefferson’s guidelines. The results indicated that the religious exclamations are noticeably context-bound, and Iranian Shiites regularly yet variously employed them to perform different speech acts and communicative functions, which are essentially different from the original religious purposes of the exclamations. In addition, it was revealed that the exclamations with different illocutionary meanings are clearly marked in terms of suprasegmental features. The findings are showcased and discussed pointing to illustrative conversations.

English Lingua Franca: New Parameters for the Teaching (and Testing) of English Pronunciation?

The recent (2018) Companion Volume to the Common European Framework offers an overhaul of many of the scales of descriptors, including, notably, phonology. A single, skeletal, scale for ‘phonological control’ is replaced by three scales, describing overall control, sound articulation, and prosodic features. In each of these, the focus has become intelligibility, rather than proximity to a native speaker accent. In this article I examine the development of pronunciation teaching since the communicative revolution, and the rise of English as a lingua franca (ELF) in which intelligibility is crucial. The article concludes with a reflection on how (if at all) the revised framework could inform an ELF aware assessment of pronunciation.

Feedback and Feedforward Auditory-Motor Processes for Voice and Articulation in Parkinson's Disease.

Unexpected and sustained manipulations of auditory feedback during speech production result in "reflexive" and "adaptive" responses, which can shed light on feedback and feedforward auditory-motor control processes, respectively. Persons with Parkinson's disease (PwPD) have shown aberrant reflexive and adaptive responses, but responses appear to differ for control of vocal and articulatory features. However, these responses have not been examined for both voice and articulation in the same speakers and with respect to auditory acuity and functional speech outcomes (speech intelligibility and naturalness). Here, 28 PwPD on their typical dopaminergic medication schedule and 28 age-, sex-, and hearing-matched controls completed tasks yielding reflexive and adaptive responses as well as auditory acuity for both vocal and articulatory features. No group differences were found for any measures of auditory-motor control, conflicting with prior findings in PwPD while off medication. Auditory-motor measures were also compared with listener ratings of speech function: first formant frequency acuity was related to speech intelligibility, whereas adaptive responses to vocal fundamental frequency manipulations were related to speech naturalness. These results support that auditory-motor processes for both voice and articulatory features are intact for PwPD receiving medication. This work is also the first to suggest associations between measures of auditory-motor control and speech intelligibility and naturalness.

Classification of emotions and evaluation of customer satisfaction from speech in real world acoustic environments

This paper focuses on finding suitable features to robustly recognize emotions and evaluate customer satisfaction (CS) from speech in real acoustic scenarios. The classification of emotions is based on standard and well-known corpora and the evaluation of CS is based on recordings of real opinions given by customers about the received service during phone calls with call-center agents. The feature sets considered in this study include two speaker models, namely x-vectors and i-vectors, and also the well known feature set introduced in the Interspeech 2010 Paralinguistics Challenge (I2010PC). Additionally, we introduce the use of phonation, articulation and prosody features extracted with the DisVoice framework as alternative feature sets to robustly model emotions and CS from speech. The results indicate that the I2010PC feature set is the best approach to classify emotions in the standard databases typically used in the literature. When considering the recordings collected in the call-center, without any control over the acoustic conditions, the best results are those that include our articulation features. The I2010PC feature set includes 1584 measures while the articulation approach only includes 488 measures. We think that the proposed approach is more suitable for real-world applications where the acoustic conditions are not controlled and also it is potentially more convenient for industrial applications.

Articulatory features of phonemes pattern to iconic meanings: evidence from cross-linguistic ideophones

Abstract Iconic words are supposed to exhibit imitative relationships between their linguistic forms and their referents. Many studies have worked to pinpoint sound-to-meaning correspondences for ideophones from different languages. The correspondence patterns show similarities across languages, but what makes such language-specific correspondences universal, as iconicity claims to be, remains unclear. This could be due to a lack of consensus on how to describe and test the perceptuo-motor affordances that make an iconic word feel imitative to speakers. We created and analysed a database of 1,860 ideophones across 13 languages, and found that seven articulatory features, physiologically accessible to all spoken language users, pattern according to semantic features of ideophones. Our findings pave the way for future research to utilize articulatory properties as a means to test and explain how iconicity is encoded in spoken language. The perspective taken here fits in with ongoing research of embodiment, motivation, and iconicity research, three major strands of research within Cognitive Linguistics. The results support that there is a degree of unity between the concepts of imitative communication and the spoken forms through cross-domain mappings, which involve physical articulatory movement.

ACADEMICIAN NIGMET SAURANBAYEV IS A PHONETICS SCIENTIST

The article is dedicated to the scientific heritage of the outstanding linguist, academician N. Sauranbayev. The scientist is especially known as a specialist in the syntax of the Kazakh language. His works on phonetics and phonology have long gone beyond the scope of research. The article discusses the position of the scientist on segmental phonetics. Differential features of phonemes, their classification, articulatory characteristics of vowels and consonants, the ratio of sounds and letters, accents and syllables are analyzed.

Rethinking the phonetics of baby-talk: Differences across Canada and Vanuatu in the articulation of mothers' speech to infants.

Infant-directed speech (IDS) is phonetically distinct from adult-directed speech (ADS): It is typically considered to have special prosody-like higher pitch and slower speaking rates-as well as unique speech sound properties, for example, more breathy, hyperarticulated, and/or variable consonant and vowel articulation. These phonetic features are widely observed in the IDS of caregivers from urbanized contexts who speak a handful of very well-researched languages. Yet studies with more diverse socio-cultural and linguistic samples show that this "typical" IDS prosody is not consistently observed across cultures. We extended cross-cultural work by examining IDS speech segment articulation, which-like prosody-is also thought to be a characteristic phonetic feature of IDS that might aid speech and language development. Here we asked whether IDS vowels have different articulatory features compared to ADS vowels in two distinct linguistic and socio-cultural contexts: urban English-speaking Canadian mothers, and rural Lenakel- and Southwest Tanna-speaking ni-Vanuatu mothers (n=57, 20-46 years of age). Replicating prior work, Canadian mothers had more variable vowels in IDS compared to ADS, but also did not show clear register differences for breathiness or hyperarticulation. Vowels spoken by ni-Vanuatu mothers showed very distinct articulatory tendencies, using less variable (and less breathy) IDS vowels. Along with other work showing diversity in IDS phonetics across populations, this paper suggests that any understanding of how IDS might aid speech and language development are best examined through a culturally- and linguistically-specific lens.

Emotion recognition with speech articulatory coordination features

Mental health illnesses like Major Depressive Disorder and Schizophrenia affect the coordination between articulatory gestures in speech production. Coordination features derived from Vocal tract variables (TVs) predicted by a speech inversion system can quantify the changes in articulatory gestures and have proven to be effective in the classification of mental health disorders. In this study we use data from the IEMOCAP (acted emotions) and MSP Podcast (natural emotions) datasets to understand how coordination features extracted from TVs can be used to capture changes between different emotions for the first time. We compared the eigenspectra extracted from channel delay correlation matrices for Angry, Sad and Happy emotions with respect to the “Neutral” emotion. Across both the datasets, it was observed that the “Sad” emotion follows a pattern suggesting simpler articulatory coordination while the “Angry” emotion follows the opposite showing signs of complex articulatory coordination. For the majority of subjects, the ‘Happy’ emotion follows a complex articulatory coordination pattern, but has significant confusion with “Neutral” emotion. We trained a Convolutional Neural Network with the coordination features as inputs to perform emotion classification. A detailed interpretation of the differences in eigenspectra and the results of the classification experiments will be discussed.