Representation Of Tone Research Articles

Learning and consolidating the contextualized contour representations of tones from F0 sequences and durational variations via transformers

Many speech characteristics, including conventional acoustic features such as mel frequency cepstrum coefficients and mel-spectrograms, as well as pre-trained contextualized acoustic representations such as wav2vec2.0, are used in a deep neural network or successfully fine-tuned with a connectionist temporal classification for Mandarin tone classification. In this study, the authors propose a transformer-based tone classification architecture, TNet-Full, which uses estimated fundamental frequency (F0) values and aligned boundary information on syllables and words. Key components of the model framework are the contour encoder and rhythm encoder, as well as the cross-attention between contours and rhythms established in the interaction encoder. Using contextual tonal contours as a reference, as well as rhythmic information derived from duration variations to consolidate more on contour representations for tone recognition, TNet-Full achieves absolute performance improvements of 24.4% for read speech (from 71.4% to 95.8%) and 6.3% for conversational speech (from 52.1% to 58.4%) when compared to a naive, simple baseline transformer, TNet-base. The relative improvements are 34.2% and 12.1%. As humans perceive tones, contour abstractions of tones can only be derived from F0 sequences, and tone recognition would be improved if syllable temporal organization was stable and predictable instead of fluctuating as seen in conversations.

Diversifying Skin Tone Representation in Medical School Dermatology Curriculum

AbstractThe lack of diversity in medical education contributes to disparities in dermatologic health outcomes, particularly for patients of color. This study evaluates second-year medical students’ comfort in diagnosing conditions in skin of color (SOC) and explores strategies to improve their confidence. We hypothesize that a curriculum enriched with SOC representation will enhance diagnostic skills. Surveys from Fall 2020 and Fall 2022 showed increased student comfort in diagnosing SOC. Fall 2022 students, exposed to more diverse dermatologic photos, demonstrated greater confidence. These findings underscore the necessity of integrating SOC content into medical curricula to address healthcare disparities.

53322 The Representation of Skin Tones in Google Images of Skin Cancers

53322 The Representation of Skin Tones in Google Images of Skin Cancers

Incorporating Frequency Effects in the Lexical Access of Mandarin Tone 3 Sandhi.

Mandarin tone 3 sandhi refers to the phenomenon whereby a tone 3 syllable changes to a tone 2 when followed by another tone 3. This phonological process creates a deviation between the tonal forms realized at morphemic (/tone3-tone3/) and word ([tone2-tone3]) levels, posing questions in terms of how disyllabic tone 3 sandhi words are represented and accessed. The current study conducted three cross-modal lexical decision priming experiments to investigate this issue. Experiment 1 manipulated the frequencies of the initial morpheme and whole word, showing that the higher initial-character frequency against the whole word gives stronger activation to the underlying representation and the lower frequency of the initial character leads to stronger activation of the surface tone. Experiments 2 and 3 operationalized the relative frequency of the initial tone 3 morpheme's realization as a sandhi tone, finding that the competition between the two tonal realizations also influences how T3 sandhi words are accessed. Specifically, the more frequently the T3 morpheme surfaces as a T2 allomorph, the less activated the underlying representation becomes in the mental lexicon. Our results indicate a complex interplay between morpheme, word, and the associated tonal representations in the mental lexicon and that these factors co-determine the lexical access of tone 3 sandhi.

781 Visual Representation of Race and Skin Tones in North American Burn Literature

781 Visual Representation of Race and Skin Tones in North American Burn Literature

Diversity and complexity of arthropod references in haiku.

Haiku are short poems, each composed of about 10 words, that typically describe moments in nature. People have written haiku since at least the 17th century, and the medium continues to be popular with poets, amateurs, educators, and students. Collectively, these poems represent an opportunity to understand which aspects of nature-e.g., which taxa and biological traits-resonate with humans and whether there are temporal trends in their representation or the emotions associated with these moments. We tested this potential using a mix of linguistic and biological methods, in analyses of nearly 4,000 haiku that reference arthropods. We documented the taxa and the life history traits represented in these poems and how they changed over time. We also analyzed the poems for emotion and tone. Our results reveal a mix of predictable trends and compelling surprises, each of which stand to potentially inform engagement strategies. At least 99 families of arthropods, in 28 orders, are represented in these haiku. The eight most commonly referenced taxa, from highest to lowest number of references, include: Lepidoptera, Hymenoptera, Diptera, Coleoptera, Araneae, Orthoptera, Hemiptera, and Odonata. Several common, conspicuous orders were never referenced, including Trichoptera, Plecoptera, and Megaloptera. The most commonly referenced traits relate to ecology (especially habitat, phenology, time of day), behavior (especially sound production), phenotype (especially color), and locomotion (especially flight). The least common traits in haiku relate to arthropod reproduction and physiology. Our analyses revealed few obvious temporal trends in the representations of taxa, biological traits, or emotion and tone. The broader implications of these results and possible future directions are discussed.

Tone selection for cochlear implants using peak-finding and frequency bin overlap

The Advanced Bionics cochlear implant devices employ a method of allocating static frequency bins to electrodes based on the natural tonotopic organization of the cochlea. Many studies show that there is a significant frequency-to-place mismatch in many subjects, but that over time the brain adjusts to these mismatches. Also, most of the tonal information in human speech is present in the lower end of the frequency spectrum, where there are few frequency bins and electrodes available to represent this information. Here, adjustments to the frequency bin allocation algorithm used in the crowdsourced CI Hackathon code are made to allow a more accurate representation of the original signal in the lower registers. A peak-finding and selecting method is used to find significant peaks regardless of frequency bin, then bins are overlapped to allow more accurate representation of tones near the borders between channels and peaks which may be near to each other.

Long-Horizon Associative Learning Explains Human Sensitivity to Statistical and Network Structures in Auditory Sequences.

Networks are a useful mathematical tool for capturing the complexity of the world. In a previous behavioral study, we showed that human adults were sensitive to the high-level network structure underlying auditory sequences, even when presented with incomplete information. Their performance was best explained by a mathematical model compatible with associative learning principles, based on the integration of the transition probabilities between adjacent and nonadjacent elements with a memory decay. In the present study, we explored the neural correlates of this hypothesis via magnetoencephalography (MEG). Participants (N = 23, 16 females) passively listened to sequences of tones organized in a sparse community network structure comprising two communities. An early difference (∼150 ms) was observed in the brain responses to tone transitions with similar transition probability but occurring either within or between communities. This result implies a rapid and automatic encoding of the sequence structure. Using time-resolved decoding, we estimated the duration and overlap of the representation of each tone. The decoding performance exhibited exponential decay, resulting in a significant overlap between the representations of successive tones. Based on this extended decay profile, we estimated a long-horizon associative learning novelty index for each transition and found a correlation of this measure with the MEG signal. Overall, our study sheds light on the neural mechanisms underlying human sensitivity to network structures and highlights the potential role of Hebbian-like mechanisms in supporting learning at various temporal scales.

Pitch Improvement in Attentional Blink: A Study across Audiovisual Asymmetries.

Attentional blink (AB) is a phenomenon in which the perception of a second target is impaired when it appears within 200-500 ms after the first target. Sound affects an AB and is accompanied by the appearance of an asymmetry during audiovisual integration, but it is not known whether this is related to the tonal representation of sound. The aim of the present study was to investigate the effect of audiovisual asymmetry on attentional blink and whether the presentation of pitch improves the ability to detect a target during an AB that is accompanied by audiovisual asymmetry. The results showed that as the lag increased, the subject's target recognition improved and the pitch produced further improvements. These improvements exhibited a significant asymmetry across the audiovisual channel. Our findings could contribute to better utilizations of audiovisual integration resources to improve attentional transients and auditory recognition decline, which could be useful in areas such as driving and education.

Application of anatomy-based spacing of electrode contacts for achieving a uniform semitonal resolution: A novel concept in cochlear implant electrode design

Using anatomy-based fitting, we can determine the place-specific map with individualized center frequencies for each electrode contact that is a closer match to the natural pitch-place of the cochlea. The primary objective of this study is to evaluate the tonal presentation across the electrode array and to calculate the semitone difference between each adjacent pair of contacts according to their anatomy-based map. The secondary objective is to determine the distancing of the contacts that would result in an equal semitone difference with a uniform tonal presentation. A total of 167 ears were included in this retrospective study. The frequencies across the electrode arrays were found to be unequally presented. The semitonal condensations were higher in the apical inter-contact spaces compared to the basal inter-contact spaces, being 3.0–2.3 semitones/mm (Kruskal Wallis test, p < 0.000). The anatomy-based spacing of the electrode contacts was larger in the basal inter-contact spaces compared to the apical inter-contact spaces, ranging from 1.92 to 1.48 mm. In conclusion, the current electrode designs do not have uniform tonal representation throughout the electrode array. There is a more condensed tonal presentation in the apical electrodes than in the basal electrodes, resulting in a lower tonal resolution in the apical region.

Categorical Perception of Lexical Tones and Stops in Mandarin-Speaking Musicians and Nonmusicians

This study investigates the perception of Mandarin lexical tones and stops to examine the degree of overlap between music and language. Eighteen musicians and 21 nonmusicians participated in a typical categorical perception task. Results showed that musicians and nonmusicians had comparable degree of categorical perception of tones and stops. Compared to nonmusicians, musicians exhibited enhanced sensitivities to within-category lexical tone stimuli. However, this improved ability was not observed in the perception of stops. These findings imply that musical experience strengthens the acuity of subtle low-level acoustic variations between within-category lexical tone stimuli without interfering with the high-level phonological representations of lexical tones, and this facilitatory effect is selective and could not readily extend to stop consonants in native language.

Abstract PR001: Radiation dose estimates via live, optical imaging are highly dependent upon patient skin tone

Abstract Introduction: New medical technologies must be tested across diverse populations to ensure efficacy for all patients. In the last decade, a novel technique for imaging in vivo radiation dose was developed to mitigate clinical radiotherapy (RT) incidents. Named Cherenkov imaging (CI), this technology captures the inherent light emission from ionizing radiation interacting with tissue during RT and allows for spatial verification of dose across a patient surface compared to their plan. Recent efforts have been made to use CI to quantitatively assess the true dose delivered across a patient’s surface in real-time, but many patient-specific factors prevent developing a direct Cherenkov-to-dose calibration until addressed. Melanin, a known absorber of light including Cherenkov wavelengths, is the most influential. Early phantom studies showed a near 45% change in emission across a range of skin tones for the same dose. Up until now, further investigation into this effect has been limited to phantoms due to a low number of in vivo CI studies and predominantly Caucasian catchment area demographics. We present a first detailed analysis of imaging patients with a wider range of skin tones and quantifying the effect that it has on dose estimation errors. Methods: A multi-institutional partnership was developed to expand patient demographics and increase diverse skin tone representation. All CI was completed with time-gated iCMOS cameras that capture the low light emission from tissue during RT. Cherenkov images were normalized by absolute planned dose and bulk tissue heterogeneities to minimize non-linear effects outside of skin tone. Patients were initially categorized by ethnicity and their skin tones were quantified by analysis of calibrated color photography. Average Cherenkov light intensity per dose was compared across various skin tones towards developing a calibration system for quantifying dose from the corrected signals. Results: In vivo Cherenkov images exhibit a nearly 220% decrease in signal across a range of skin tones, from White to Black/African American patients, greatly affecting the Cherenkov-to-dose calibration. However, a linear correlation exists between quantitative skin luminosity and dose-normalized Cherenkov emission. Thus, quantifying skin luminosity provides a method to normalize out the effect of skin tone and minimize the in vivo dose estimation errors across diverse patient demographics. Conclusion: This study addresses the critically important issue of developing optical tools that are appropriately calibrated for all skin tones. In vivo CI must be calibrated by skin tone for dosimetric assessments. We have shown that it is not only feasible to image dark skin patients, but possible to quantify the effect that a range of skin tones has on CI for a given dose. Using this information, we can develop a skin tone normalization algorithm to minimize dose estimation errors. Future work will focus on expanding our patient dataset to even darker skin tones through active recruitment efforts in the participating clinics. Citation Format: Savannah Decker, Jacqueline Andreozzi, Diego Hernandez, Daniel Alexander, Vihan Wickramasinghe, Ibrahim Oraiqat, Erli Chen, Iman Washington, Rongxiao Zhang, Lesley Jarvis, Petr Bruza, David Gladstone, Brian Pogue. Radiation dose estimates via live, optical imaging are highly dependent upon patient skin tone [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr PR001.

Representational level matters for tone-word recognition: Evidence from form priming.

In a form priming experiment with a lexical decision task, we investigated whether the representational structure of lexical tone in lexical memory impacts spoken-word recognition in Mandarin. Target monosyllabic words were preceded by five types of primes: (1) the same real words (/lun4/-/lun4/), (2) real words with only tone contrasts (/lun2/-/lun4/), (3) unrelated real words (/pie3/-/lun4/), (4) pseudowords with only tone contrasts (*/lun3/-/lun4/), and (5) unrelated pseudowords (*/tai3/-/lun4/). We found a facilitation effect in target words with pseudoword primes that share the segmental syllable but contrast in tones (*/lun3/-/lun4/). Moreover, no evident form priming effect was observed in target words primed by real words with only tone contrasts (/lun2/-/lun4/). These results suggest that the recognition of a tone word is influenced by the representational level of tone accessed by the prime word. The distinctive priming patterns between real-word and pseudoword primes are best explained by the connectionist models of tone-word recognition, which assume a hierarchical representation of lexical tone.

Colorism in advertisements in Kenyan print newspapers: A study of the nation newspaper

Colorism is discriminatory, as it excludes dark skin tones and elevates light skin tones. Studies have established the negative effects of colorism on different groups of people. Owing to the power of advertisements to influence and shape opinions, beliefs and perceptions, this study’s purpose was to determine colorism in advertisements in Kenyan print newspapers. The research methodology encompassed the use of a case study design. The Nation print newspaper was studied because it is the most widely read newspaper in Kenya. Data was collected through the content analysis of all advertisements from January to April 2023. The study found out that light skin was the most preferred skin tone for males and females in the advertisements. However, more females than males in the advertisements were light-skinned, whereas more males than females were dark-skinned. This study concludes that colorism exists in advertisements in Kenyan print newspapers. The study recommends equal representations of color tones to mirror the composition of the Kenyan population.

Subjectively preferred octave size is resolved at the late stages of cerebral auditory processing.

Human listeners prefer octave intervals slightly above the exact 2:1 frequency ratio. To study the neural underpinnings of this subjective preference, called the octave enlargement phenomenon, we compared neural responses between exact, slightly enlarged, oversized, and compressed octaves (or their multiples). The first experiment (n= 20) focused on the N1 and P2 event-related potentials (ERPs) elicited in EEG 50-250 ms after the second tone onset during passive listening of one-octave intervals. In the second experiment (n= 20) applying four-octave intervals, musician participants actively rated the different octave types as 'low', 'good' and 'high'. The preferred slightly enlarged octave was individually determined prior to the second experiment. In both experiments, N1-P2 peak-to-peak amplitudes attenuated for the exact and slightly enlarged octave intervals compared with compressed and oversized intervals, suggesting overlapping neural representations of tones an octave (or its multiples) apart. While there were no differences between the N1-P2 amplitudes to the exact and preferred enlarged octaves, ERP amplitudes differed after 500 ms from onset of the second tone of the pair. In the multivariate pattern analysis (MVPA) of the second experiment, the different octave types were distinguishable (spatial classification across electroencephalography [EEG] channels) 200 ms after second tone onset. Temporal classification within channels suggested two separate discrimination processes peaking around 300 and 700 ms. These findings appear to be related to active listening, as no multivariate results were found in the first, passive listening experiment. The present results suggest that the subjectively preferred octave size is resolved at the late stages of auditory processing.

Statistical Learning of Chord-Transition Regularities in a Novel Equitempered Scale: An MMN Study

In music and language domains, it has been suggested that patterned transitions of sounds can be acquired implicitly through statistical learning. Previous studies have investigated the statistical learning of auditory regularities by recording early neural responses to a sequence of tones presented at high or low transition probabilities. However, it remains unclear whether the statistical learning of musical chord transitions is reflected in endogenous, regularity-dependent components of the event-related potential (ERP). The present study aimed to record the mismatch negativity (MMN) elicited by chord transitions that deviated from newly learned transitional regularities. Chords were generated in a novel 18 equal temperament pitch class scale to avoid interference from the existing tonal representations of the 12 equal temperament pitch class system. Thirty-six adults without professional musical training listened to a sequence of randomly inverted chords in which certain chords were presented with high (standard) or low (deviant) transition probabilities. An irrelevant timbre change detection task was assigned to make them attend to the sequence during the ERP recording. After that, a familiarity test was administered in which the participants were asked to choose the more familiar chord sequence out of two successive sequences. The results showed that deviant transitions elicited the MMN, although the participants could not recognize the standard transition beyond the level of chance. These findings suggest that humans can statistically learn new transitional regularities of chords in a novel musical scale, even though they did not recognize them explicitly. This study provides further evidence that music-syntactic regularities can be acquired implicitly through statistical learning.

44445 Representation of skin tones in Google image results for psoriasis

44445 Representation of skin tones in Google image results for psoriasis

Calling for wider skin tone representation in simulation-based learning

Calling for wider skin tone representation in simulation-based learning

Skin tone representation in dermatologic direct-to-consumer advertisements: a cross-sectional analysis and call to action.

Direct-to-consumer advertisements (DTCAs) in medical marketing serve as a prominent modality to deliver information to an increasingly diverse audience of consumers and increase prescription sales. In dermatology, advertisements have the potential to shape the public's opinions, aid in the understanding of skin conditions, and raise awareness of available treatments. To investigate and characterize the representation of skin tones in DTCAs. Nielsen ratings were utilized to identify the networks most watched by Black viewers in 2022. Programming on NBCUniversal, ABC, CBS, and FOX that aired in the District of Columbia, suburban Maryland, and Northern Virginia from June 2022 to July 2022 was reviewed for DTCAs. DTCAs were then analyzed to determine the skin tones of models and skin conditions depicted on models with darkly pigmented skin. Of the 106 DTCAs related to dermatologic conditions, there were 13 unique advertisements featuring 32 unique models. Four advertisements depicted the skin condition on darkly pigmented skin tones. Using the Monk Skin Tone (MST) scale to assess the 32 unique individuals, only 25% (n = 8) were rated at an MST 7 or above, and 6.25% (n = 2) were rated at an MST 10. This study has the limitation of only sampling DTCAs from Washington, District of Columbia which does not fully represent all dermatology-related DTCAs in the United States. Results of this content analysis demonstrate that the number of persons of color within dermatologic DTCAs is 23%, whereas there are 13.6% Black individuals in the 2021 US census. This suggests that DTCAs are becoming more diverse since 2018. However, findings also show that the vast majority of DTCAs do not include models with darkly pigmented skin, and there remains a lack of advertisements depicting skin disease among people of color. Given the role of DTCAs in informing and aiding patients' requests for prescription drugs, representation of all skin tones is essential for this communication to be effective, especially in the field of dermatology.

Skin Tone Analysis for Representation in Educational Materials (STAR-ED) using machine learning

Images depicting dark skin tones are significantly underrepresented in the educational materials used to teach primary care physicians and dermatologists to recognize skin diseases. This could contribute to disparities in skin disease diagnosis across different racial groups. Previously, domain experts have manually assessed textbooks to estimate the diversity in skin images. Manual assessment does not scale to many educational materials and introduces human errors. To automate this process, we present the Skin Tone Analysis for Representation in EDucational materials (STAR-ED) framework, which assesses skin tone representation in medical education materials using machine learning. Given a document (e.g., a textbook in .pdf), STAR-ED applies content parsing to extract text, images, and table entities in a structured format. Next, it identifies images containing skin, segments the skin-containing portions of those images, and estimates the skin tone using machine learning. STAR-ED was developed using the Fitzpatrick17k dataset. We then externally tested STAR-ED on four commonly used medical textbooks. Results show strong performance in detecting skin images (0.96 ± 0.02 AUROC and 0.90 ± 0.06 F1 score) and classifying skin tones (0.87 ± 0.01 AUROC and 0.91 ± 0.00 F1 score). STAR-ED quantifies the imbalanced representation of skin tones in four medical textbooks: brown and black skin tones (Fitzpatrick V-VI) images constitute only 10.5% of all skin images. We envision this technology as a tool for medical educators, publishers, and practitioners to assess skin tone diversity in their educational materials.