MyMedR

Displaying all 12 publications

Abstract:

Sort:

Fulltext Code-mixing unveiled: Enhancing the hate speech detection in Arabic dialect tweets using machine learning models

Alhazmi A, Mahmud R, Idris N, Mohamed Abo ME, Eke CI

PLoS One, 2024;19(7):e0305657.
PMID: 39018339 DOI: 10.1371/journal.pone.0305657

Technological developments over the past few decades have changed the way people communicate, with platforms like social media and blogs becoming vital channels for international conversation. Even though hate speech is vigorously suppressed on social media, it is still a concern that needs to be constantly recognized and observed. The Arabic language poses particular difficulties in the detection of hate speech, despite the considerable efforts made in this area for English-language social media content. Arabic calls for particular consideration when it comes to hate speech detection because of its many dialects and linguistic nuances. Another degree of complication is added by the widespread practice of "code-mixing," in which users merge various languages smoothly. Recognizing this research vacuum, the study aims to close it by examining how well machine learning models containing variation features can detect hate speech, especially when it comes to Arabic tweets featuring code-mixing. Therefore, the objective of this study is to assess and compare the effectiveness of different features and machine learning models for hate speech detection on Arabic hate speech and code-mixing hate speech datasets. To achieve the objectives, the methodology used includes data collection, data pre-processing, feature extraction, the construction of classification models, and the evaluation of the constructed classification models. The findings from the analysis revealed that the TF-IDF feature, when employed with the SGD model, attained the highest accuracy, reaching 98.21%. Subsequently, these results were contrasted with outcomes from three existing studies, and the proposed method outperformed them, underscoring the significance of the proposed method. Consequently, our study carries practical implications and serves as a foundational exploration in the realm of automated hate speech detection in text.

Matched MeSH terms: Speech/physiology
Fulltext A magnetic resonance imaging study on the articulatory and acoustic speech parameters of Malay vowels

Zourmand A, Mirhassani SM, Ting HN, Bux SI, Ng KH, Bilgen M, et al.

Biomed Eng Online, 2014;13:103.
PMID: 25060583 DOI: 10.1186/1475-925X-13-103

The phonetic properties of six Malay vowels are investigated using magnetic resonance imaging (MRI) to visualize the vocal tract in order to obtain dynamic articulatory parameters during speech production. To resolve image blurring due to the tongue movement during the scanning process, a method based on active contour extraction is used to track tongue contours. The proposed method efficiently tracks tongue contours despite the partial blurring of MRI images. Consequently, the articulatory parameters that are effectively measured as tongue movement is observed, and the specific shape of the tongue and its position for all six uttered Malay vowels are determined.Speech rehabilitation procedure demands some kind of visual perceivable prototype of speech articulation. To investigate the validity of the measured articulatory parameters based on acoustic theory of speech production, an acoustic analysis based on the uttered vowels by subjects has been performed. As the acoustic speech and articulatory parameters of uttered speech were examined, a correlation between formant frequencies and articulatory parameters was observed. The experiments reported a positive correlation between the constriction location of the tongue body and the first formant frequency, as well as a negative correlation between the constriction location of the tongue tip and the second formant frequency. The results demonstrate that the proposed method is an effective tool for the dynamic study of speech production.

Matched MeSH terms: Speech/physiology*
Vocal fold vibratory characteristics in normal female speakers from high-speed digital imaging

Ahmad K, Yan Y, Bless DM

J Voice, 2012 Mar;26(2):239-53.
PMID: 21621975 DOI: 10.1016/j.jvoice.2011.02.001

The purpose of the study was to investigate relationships between vocal fold vibrations and voice quality. Laryngeal images obtained from high-speed digital imaging (HSDI) were examined for their open-closed timing characteristics and perturbation values. A customized software delineated the glottal edges and used the Hilbert transform-based method of analysis to provide objective quantification of glottal perturbation. Overlay tracings of the transformed glottal cycles provided visual patterns on the overall vibratory dynamics. In this paper, we described the use of this method in looking at vibratory characteristics of a group of young female speakers (N=23). We found that, females with no voice complaints and who had been perceived to have normal voices were not a homogeneous group in terms of their glottal vibratory patterns during phonation. Their vibratory patterns showed characteristics similar to exemplar voices targeted to be clear (50%), pressed (27%), breathy (15%), or a mixed quality (8%). Perturbation range in terms of cycle-to-cycle frequency and amplitude was small and did not discriminate patterns. All these patterns yielded perceptually normal voices suggesting that in normal young speakers, the level of perturbation may be more important to the judgment than the actual pattern of closure.

Matched MeSH terms: Speech/physiology
Human behavior state profile mapping based on recalibrated speech affective space model

Kamaruddin N, Wahab A

Annu Int Conf IEEE Eng Med Biol Soc, 2012;2012:2021-4.
PMID: 23366315 DOI: 10.1109/EMBC.2012.6346354

People typically associate health with only physical health. However, health is also interconnected to mental and emotional health. People who are emotionally healthy are in control of their behaviors and experience better quality of life. Hence, understanding human behavior is very important in ensuring the complete understanding of one's holistic health. In this paper, we attempt to map human behavior state (HBS) profiles onto recalibrated speech affective space model (rSASM). Such an approach is derived from hypotheses that: 1) Behavior is influenced by emotion, 2) Emotion can be quantified through speech, 3) Emotion is dynamic and changes over time and 4) the emotion conveyance is conditioned by culture. Empirical results illustrated that the proposed approach can complement other types of behavior analysis in such a way that it offers more explanatory components from the perspective of emotion primitives (valence and arousal). Four different driving HBS; namely: distracted, laughing, sleepy and normal are profiled onto the rSASM to visualize the correlation between HBS and emotion. This approach can be incorporated in the future behavior analysis to envisage better performance.

Matched MeSH terms: Speech/physiology*
Speech stimuli and nasalance scores for the assessment of resonance in in Mandarin speaking Malaysian children

Mohd Ibrahim H, Lim HW, Ahmad Rusli Y, Lim CT

Clin Linguist Phon, 2020 06 02;34(6):554-565.
PMID: 31537131 DOI: 10.1080/02699206.2019.1668480

This study was designed to develop language-specific stimuli for the assessment of resonance and to obtain nasalance scores using the newly developed speech stimuli in Mandarin. Gender and age influences on nasalance scores for each of the stimulus were also examined. Participants recruited were typically developing Mandarin-speaking ethnic Chinese children aged 6;00-7;11 growing up in Malaysia. Perceptual ratings of nasality were made based on the GOS.SP.ASS.'98 (revised) for children while nasalance scores were recorded for each stimulus using the Nasometer II (Model 6400). Fifty Mandarin-speaking children (24 males and 26 females) were recruited. None of the participants were perceived with abnormal nasality on the three stimuli. The mean nasalance scores for the Mandarin stimuli were 16.08% (SD = 2.57, 95% CI = 15.35-16.81) for the Oral passage, 25.20% (SD = 3.63, 95% CI = 24.17-26.23) for the Oral-Nasal passage and 55.44% (SD = 4.17, 95% CI = 54.25-56.63) for the Nasal passage. No significant age- and gender-related differences were observed for all the three stimuli. This is the first set of Mandarin stimuli and nasalance norms for Mandarin-speaking children in Malaysia. The influence of phonetic content on nasalance is supported. Findings call for language-specific normative nasalance data and careful selection of stimuli for the assessment of resonance.

Matched MeSH terms: Speech/physiology*
Animacy-Based Accessibility and Competition in Relative Clause Production in Hindi and Malayalam

Perera CK, Srivastava AK

J Psycholinguist Res, 2016 Aug;45(4):915-30.
PMID: 26108301 DOI: 10.1007/s10936-015-9384-0

Two studies investigated interaction of animacy-based accessibility and competition processes with language specific constraints in shaping production preferences. Relative clause elicitation tasks (Gennari et al. in Cogn Psychol 65:141-176, 2012) were performed by two groups with 40 participants in each. Significantly more passives were produced with animate question/focus and assigned subject function to them in Malayalam, while the difference was insignificant in Hindi with animacy not equating with grammatical function assignment. Both languages produced active objects with OSV order significantly more with animate question/focus. This indicates animacy effect on word order. Animacy also influenced the expression of the agent phase in both languages as a tendency was found to omit the agent by-phase or to delay its appearance, when the two animate nouns entering into the structure are highly similar. The results point to language specific constraints in relative clause production and variability in the role of animacy-based retrieval order across languages.

Matched MeSH terms: Speech/physiology*
Fulltext Speech token detection and discrimination in individual infants using functional near-infrared spectroscopy

Mao D, Wunderlich J, Savkovic B, Jeffreys E, Nicholls N, Lee OW, et al.

Sci Rep, 2021 Dec 14;11(1):24006.
PMID: 34907273 DOI: 10.1038/s41598-021-03595-z

Speech detection and discrimination ability are important measures of hearing ability that may inform crucial audiological intervention decisions for individuals with a hearing impairment. However, behavioral assessment of speech discrimination can be difficult and inaccurate in infants, prompting the need for an objective measure of speech detection and discrimination ability. In this study, the authors used functional near-infrared spectroscopy (fNIRS) as the objective measure. Twenty-three infants, 2 to 10 months of age participated, all of whom had passed newborn hearing screening or diagnostic audiology testing. They were presented with speech tokens at a comfortable listening level in a natural sleep state using a habituation/dishabituation paradigm. The authors hypothesized that fNIRS responses to speech token detection as well as speech token contrast discrimination could be measured in individual infants. The authors found significant fNIRS responses to speech detection in 87% of tested infants (false positive rate 0%), as well as to speech discrimination in 35% of tested infants (false positive rate 9%). The results show initial promise for the use of fNIRS as an objective clinical tool for measuring infant speech detection and discrimination ability; the authors highlight the further optimizations of test procedures and analysis techniques that would be required to improve accuracy and reliability to levels needed for clinical decision-making.

Matched MeSH terms: Speech/physiology*
The speech evoked potential in normal subjects and patients with cerebral hemispheric lesions

Pratap RC

Clin Neurol Neurosurg, 1987;89(4):237-42.
PMID: 3690926

The present study deals with observations on the "speech evoked potential"-a late positive potential evoked by word repetition. These potentials, evoked by "silent" repetition of polysyllabic words, were averaged and recorded from the scalp overlying the inferior frontal regions on both sides in 20 normal healthy subjects of ages ranging from 13-58 years. The potential had a triphasic negative, positive, negative morphology and was present over both hemispheres in left as well as right handed subjects. The main positive deflection and mean latencies of 219.2 msec and 221.6 msec and mean amplitude of 6.2 muv and 6.5 muv respectively on the left and right sides. Though there were interindividual variations in latency, amplitude and morphology, there was a high degree of intraindividual similarity and reproducibility in subjects. The variations in these parameters with age, sex and handedness are discussed. In 10 patients with cerebral lesions, the evoked potential was normal in 5 cases with right frontal lesions and showed abnormalities in 3 of 5 cases with left frontal lesions. The speech evoked potential may be useful in the further study of electrical correlates of speech output in speech disorders.

Matched MeSH terms: Speech/physiology*
Fulltext Gender differences in binaural speech-evoked auditory brainstem response: are they clinically significant?

Jalaei B, Azmi MHAM, Zakaria MN

Braz J Otorhinolaryngol, 2018 05 17;85(4):486-493.
PMID: 29858160 DOI: 10.1016/j.bjorl.2018.04.005

INTRODUCTION: Binaurally evoked auditory evoked potentials have good diagnostic values when testing subjects with central auditory deficits. The literature on speech-evoked auditory brainstem response evoked by binaural stimulation is in fact limited. Gender disparities in speech-evoked auditory brainstem response results have been consistently noted but the magnitude of gender difference has not been reported.
OBJECTIVE: The present study aimed to compare the magnitude of gender difference in speech-evoked auditory brainstem response results between monaural and binaural stimulations.
METHODS: A total of 34 healthy Asian adults aged 19-30 years participated in this comparative study. Eighteen of them were females (mean age=23.6±2.3 years) and the remaining sixteen were males (mean age=22.0±2.3 years). For each subject, speech-evoked auditory brainstem response was recorded with the synthesized syllable /da/ presented monaurally and binaurally.
RESULTS: While latencies were not affected (p>0.05), the binaural stimulation produced statistically higher speech-evoked auditory brainstem response amplitudes than the monaural stimulation (p<0.05). As revealed by large effect sizes (d>0.80), substantive gender differences were noted in most of speech-evoked auditory brainstem response peaks for both stimulation modes.
CONCLUSION: The magnitude of gender difference between the two stimulation modes revealed some distinct patterns. Based on these clinically significant results, gender-specific normative data are highly recommended when using speech-evoked auditory brainstem response for clinical and future applications. The preliminary normative data provided in the present study can serve as the reference for future studies on this test among Asian adults.

Matched MeSH terms: Speech/physiology*
Functional brain imaging of walking while talking - An fNIRS study

Metzger FG, Ehlis AC, Haeussinger FB, Schneeweiss P, Hudak J, Fallgatter AJ, et al.

Neuroscience, 2017 02 20;343:85-93.
PMID: 27915210 DOI: 10.1016/j.neuroscience.2016.11.032

Since functional imaging of whole body movements is not feasible with functional magnetic resonance imaging (fMRI), the present study presents in vivo functional near-infrared spectroscopy (fNIRS) as a suitable technique to measure body movement effects on fronto-temporo-parietal cortical activation in single- and dual-task paradigms. Previous fNIRS applications in studies addressing whole body movements were typically limited to the assessment of prefrontal brain areas. The current study investigated brain activation in the frontal, temporal and parietal cortex of both hemispheres using functional near-infrared spectroscopy (fNIRS) with two large 4×4 probe-sets with 24 channels each during single and dual gait tasks. 12 young healthy adults were measured using fNIRS walking on a treadmill: the participants performed two single-task (ST) paradigms (walking at different speeds, i.e. 3 and 5km/h) and a dual task (DT) paradigm where a verbal fluency task (VFT) had to be executed while walking at 3km/h. The results show an increase of activation in Broca's area during the more advanced conditions (ST 5km/h vs. ST 3km/h, DT vs. ST 3km/h, DT vs. 5km/h), while the corresponding area on the right hemisphere was also activated. DT paradigms including a cognitive task in conjunction with whole body movements elicit wide-spread cortical activation patterns across fronto-temporo-parietal areas. An elaborate assessment of these activation patterns requires more extensive fNIRS assessments than the traditional prefrontal investigations, e.g. as performed with portable fNIRS devices.

Matched MeSH terms: Speech/physiology*
Test-retest reliability of speech-evoked auditory brainstem response in healthy children at a low sensation level

Zakaria MN, Jalaei B

Int J Pediatr Otorhinolaryngol, 2017 Nov;102:28-31.
PMID: 29106871 DOI: 10.1016/j.ijporl.2017.08.033

OBJECTIVE: Auditory brainstem responses evoked by complex stimuli such as speech syllables have been studied in normal subjects and subjects with compromised auditory functions. The stability of speech-evoked auditory brainstem response (speech-ABR) when tested over time has been reported but the literature is limited. The present study was carried out to determine the test-retest reliability of speech-ABR in healthy children at a low sensation level.
METHODS: Seventeen healthy children (6 boys, 11 girls) aged from 5 to 9 years (mean = 6.8 ± 3.3 years) were tested in two sessions separated by a 3-month period. The stimulus used was a 40-ms syllable /da/ presented at 30 dB sensation level.
RESULTS: As revealed by pair t-test and intra-class correlation (ICC) analyses, peak latencies, peak amplitudes and composite onset measures of speech-ABR were found to be highly replicable. Compared to other parameters, higher ICC values were noted for peak latencies of speech-ABR.
CONCLUSION: The present study was the first to report the test-retest reliability of speech-ABR recorded at low stimulation levels in healthy children. Due to its good stability, it can be used as an objective indicator for assessing the effectiveness of auditory rehabilitation in hearing-impaired children in future studies.

Matched MeSH terms: Speech/physiology
Bulbar signs in normal population

Lim KS, Hew YC, Lau HK, Lim TS, Tan CT

Can J Neurol Sci, 2009 Jan;36(1):60-4.
PMID: 19294890

BACKGROUND AND OBJECTIVES: There is lack of published data on bulbar signs among the healthy population. This study aims to determine the range of normality of bulbar signs particularly among the elderly.
METHODS: Systemic examination of bulbar signs was carried out according to a predetermined protocol on a cohort of young and elderly healthy subjects.
RESULTS: A total of 206 subjects were recruited in the study, 104 young adults with mean age of 20 years, and 102 elderly with mean age of 73 years. Uvula deviation was seen in 28 (26.9%) young subjects and 22 (21.6%) elderly. Irregular tongue border was seen in 17 subjects, unilateral in 4 subjects. Fourteen (6.8%) subjects had deviation on tongue protrusion. Occasional tremor of tongue on protrusion is common in both young and old. Persistent (severe) tongue tremor on protrusion was seen in 18.6% of the elderly, and 4.8% of the young. None of the subjects had tremor of tongue at rest. In gag reflex, absence of gagging response was common in elderly, seen in two thirds of the subjects on stimulation of the posterior pharyngeal wall. However, all the subjects had uvular movement. Habituation or suppression of gagging response was seen in close to 90% of young males.
CONCLUSION: There is wide range of normality in bulbar signs in normal population, particularly among the elderly.

Matched MeSH terms: Speech/physiology

Filters

Please provide feedback to Administrator (afdal@afpm.org.my)

External Links