Phoneme categorization relying solely on high-frequency energy

Vitela, A. Davi; Monson, Brian B.; Lotto, Andrew J.

doi:10.1121/1.4903917

Cited by 40 publications

(26 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand MFCC #11 captures information in high frequency [9.5kHz -12.6kHz]. Recently researchers, supported by the advancements in equipment technology that captures a broader spectrum [47,48], have shown that there is perceptually relevant information on high frequency speech, affecting speech intelligibility. Both [47,49] concluded that high frequency speech characteristics are different in dysphonic versus control subjects, suggesting that the hoarseness characteristic of PD subjects, or in our case, the difference in hoarseness between "ON" and "OFF" states is what we are capturing with high coefficients of the MFCC.…”

Section: Discussionmentioning

confidence: 99%

Speech-based identification of L-DOPA ON/OFF state in Parkinson’s Disease subjects

Norel

Agurto

Rice

et al. 2018

Preprint

View full text Add to dashboard Cite

Background: Parkinson's disease patients (PDP) are evaluated using the unified Parkinson's disease rating scale (UP-DRS) to follow the longitudinal course of the disease. UP-DRS evaluation is performed by a neurologist, and hence its use is limited in the evaluation of short-term (daily) fluctuations. Subjects taking L-DOPA as part of treatment to reduce symptoms exhibit motor fluctuations as a common complication. Objectives: The aim of the study is to assess the use of speech analysis as a proxy to continuously monitor PDP medication state. Methods: We combine acoustic, prosody, and semantic features to characterize three speech tasks (picture description, reverse counting and diadochokinetic rate) of 25 PDP evaluated under different medication states: "ON" and "OFF" L-DOPA. Results: Classification of medication states using features extracted from audio recordings results in cross-validated accuracy rates of 0.88, 0.84 and 0.71 for the picture description, reverse counting and diadochokinetic rate tasks, respectively. When adding feature selection and semantic features, the accuracy rates increase to 1.00, 0.96 and 0.83 respectively; thus reaching very high classification accuracy on 3 different tasks. Conclusions:We show that speech-based features are highly predictive

show abstract

Section: Discussionmentioning

confidence: 99%

Speech-based identification of L-DOPA ON/OFF state in Parkinson’s Disease subjects

Norel

Agurto

Rice

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…The influence of these statistically-defined tone sequences on subsequent speech categorization suggests that listeners may keep a running estimate of the distributionally-defined LTAS across both speech and nonspeech sounds and encode subsequent sounds relative to, and contrastively with, these running averages [ 21 , 22 , 25 ]. Spectral contrast as a function of the LTAS may be an effective, domain-general process contributing to accommodation of talker differences across speech [ 21 – 24 , 26 , 29 – 33 ], including normalization of the sort described by Ladefoged and Broadbent [ 28 ], accommodation of individual differences in overall voice pitch that impact Mandarin lexical tone realization [ 24 ], the ability to adapt to a speaker’s style [casual vs. careful; 32 ] and the inability to adapt to some particular voice changes [ 34 ].…”

Section: Introductionmentioning

confidence: 99%

Short-term adaptation to sound statistics is unimpaired in developmental dyslexia

Gabay

Holt

2018

PLoS ONE

View full text Add to dashboard Cite

Developmental dyslexia is presumed to arise from phonological impairments. Accordingly, people with dyslexia show speech perception deficits taken as indication of impoverished phonological representations. However, the nature of speech perception deficits in those with dyslexia remains elusive. Specifically, there is no agreement as to whether speech perception deficits arise from speech-specific processing impairments, or from general auditory impairments that might be either specific to temporal processing or more general. Recent studies show that general auditory referents such as Long Term Average Spectrum (LTAS, the distribution of acoustic energy across the duration of a sound sequence) affect speech perception. Here we examine the impact of preceding target sounds’ LTAS on phoneme categorization to assess the nature of putative general auditory impairments associated with dyslexia. Dyslexic and typical listeners categorized speech targets varying perceptually from /ga/-/da/ preceded by speech and nonspeech tone contexts varying. Results revealed a spectrally contrastive influence of the preceding context LTAS on speech categorization, with a larger magnitude effect for nonspeech compared to speech precursors. Importantly, there was no difference in the presence or magnitude of the effects across dyslexia and control groups. These results demonstrate an aspect of general auditory processing that is spared in dyslexia, available to support phonemic processing when speech is presented in context.

show abstract

“…Age-related hearing loss, noise-induced hearing loss (NIHL), and ototoxicity may be observed as a highfrequency hearing loss that gradually progresses toward lower frequencies (Durrant et al, 2009;Seddon et al, 2012;Mehrparvar et al, 2014). The acoustic energy of extended high frequencies (EHFs) plays an important role in speech perception, especially in the presence of background noise (Rodríguez Valiente et al, 2014;Vitela et al, 2014;Vlaming et al, 2014). Despite this, the gradual change in hearing sensitivity may, initially, go unnoticed, as hearing perception is dominated by low-frequency hearing (Vlaming et al, 2014).…”

mentioning

confidence: 99%

Extended High-Frequency Smartphone Audiometry: Validity and Reliability

et al. 2018

View full text Add to dashboard Cite

Background: Extended high-frequency (EHF) audiometry (8?16 kHz) has an important role in audiological assessments such as ototoxicity monitoring, and for speech recognition and localization. Accurate and reliable EHF testing with smartphone technologies has the potential to provide more affordable and accessible hearing-care services, especially in underserved contexts. Purpose: To determine the accuracy and test?retest reliability of EHF audiometry with a smartphone application, using calibrated headphones. Research Design: Air-conduction thresholds (8?16 kHz) and test?retest reproducibility, recorded with conventional audiometry (CA) and smartphone audiometry (SA), using audiometric (Sennheiser HDA 300 circumaural) and nonstandard audiometric (Sennheiser HD202 II supra-aural) headphones, were compared in a repeated-measures design. Study Sample: A total of 61 participants (122 ears) were included in the study. Of these, 24 were adults attending a tuberculosis clinic (mean age = 36.8, standard deviation [SD] = 14.2 yr; 48% female) and 37 were adolescents and young adults recruited from a prospective students program (mean age = 17.6, SD = 3.2 yr; 76% female). Of these, 22.3% (n = 326) of EHF thresholds were ?25 dB HL. Data Analysis: Threshold comparisons were made between CA and SA, with audiometric headphones and nonstandard audiometric headphones. A paired samples t-test was used for comparison of threshold correspondence between conventional and smartphone thresholds, and test?retest reproducibility of smartphone thresholds. Results: Conventional thresholds corresponded with smartphone thresholds at the lowest intensity (10 dB HL), using audiometric and nonstandard audiometric headphones in 59.4% and 57.6% of cases, respectively. Conventional thresholds (exceeding 10 dB HL) corresponded within 10 dB or less, with smartphone thresholds in 82.9% of cases using audiometric headphones and 84.1% of cases using nonstandard audiometric headphones. There was no significant difference between CA and SA, using audiometric headphones across all frequencies (p > 0.05). Test?retest comparison also showed no significant differences between conditions (p > 0.05). Smartphone test?retest thresholds corresponded within 10 dB or less in 86.7% and 93.4% of cases using audiometric and nonstandard audiometric headphones, respectively. Conclusions: EHF smartphone testing with calibrated headphones can provide an accurate and reliable option for affordable mobile audiometry. The validity of EHF smartphone testing outside a sound booth as a cost-effective and readily available option to detect high-frequency hearing loss in community-based settings should be established.

show abstract

Phoneme categorization relying solely on high-frequency energy

Cited by 40 publications

References 16 publications

Speech-based identification of L-DOPA ON/OFF state in Parkinson’s Disease subjects

Speech-based identification of L-DOPA ON/OFF state in Parkinson’s Disease subjects

Short-term adaptation to sound statistics is unimpaired in developmental dyslexia

Extended High-Frequency Smartphone Audiometry: Validity and Reliability

Contact Info

Product

Resources

About