搜索结果：sentences

The present experiment examined whether an affective evaluation process contributes to discourse comprehension when participants silently read for comprehension, by analyzing event-related potentials (ERPs). Kunkel, Filik, Mackenzie, and Leuthold (2018) showed that when participants read about emotional events or moral transgressions while performing explicit emotional judgments, a larger late positive potential (LPP) was triggered for negative versus non-negative and immoral versus moral scenarios. They took this LPP effect to reflect the affective evaluation of incoming linguistic information with emotional content. Previous work showing implicit evaluation of morality-based materials (i.e., when participants are simply reading for comprehension) had intermixed these stimuli with emotionally-neutral world knowledge violation materials (Leuthold, Kunkel, Mackenzie, & Filik, 2015). By placing these morality materials with emotion-related materials instead, in the current study, we can more directly evaluate whether similar affective processing occurs for both types of stimuli in the absence of an explicit judgment task. Target sentences from negative versus non-negative emotional scenarios and from moral versus immoral scenarios were presented using rapid serial visual presentation. Importantly, the analysis of single-trial amplitudes using a linear mixed-effects modeling (LME) approach revealed a larger late positive potential (LPP) for negative versus non-negative and immoral versus moral scenarios, displaying a more posterior distribution for emotion versus morality condition effects. We conclude that discourse comprehension involves the implicit affective evaluation (LPP) of the emotional content of incoming linguistic information independent of the material-specific context.

Aging often reduces opportunities for everyday interaction and instrumental support, raising questions about the potential role of artificial intelligence (AI) technologies in later-life well-being. This study examined how AI voice assistants may fulfill, or fall short of, older adults' basic psychological needs through the lens of self-determination theory (SDT). Eighteen U.S.-based community-dwelling older adults (ages 66-85) participated in a two-week AI voice assistant trial and completed daily diaries and pre- and post-intervention surveys. Using mixed methods, qualitative thematic analysis and large language model-assisted content analysis, we analyzed 237 diary entries (1,512 sentences) across three SDT dimensions: autonomy, competence, and relatedness. Findings reveal voice assistants primarily supported autonomy (medication reminders, safety, emotional regulation) and competence (task accomplishment, activity maintenance), with experiences remaining largely consistent across the two-week period. Relatedness responses were the most divided: some participants valued companionship features while others found them inauthentic or uncomfortable. Despite positive daily experiences, loneliness and life satisfaction showed no significant change. These findings suggest that the benefits of AI voice assistants for older adults may be more evident in supporting everyday autonomy and competence than in addressing social isolation, and that psychological benefits of short-term use may not immediately translate into changes in global well-being outcomes.

Integrating auditory and visual information can improve intelligibility and neural tracking of the speech envelope. Speech-relevant tactile information also improves tracking, yet effects on intelligibility are mixed, possibly due to no regular exposure to speech-relevant tactile information. We used short-term audio-tactile training to advance understanding of audio-tactile integration during speech perception. 64 younger adult participants (aged 18-29) took part over five days, during which electroencephalography was recorded. Participants completed a speech-in-noise task, with audio-tactile and audio-only stimuli. They then received training with either tactile information that was congruent with sentences heard (trained group) or incongruent (pseudo-trained group). After three training sessions, participants completed the speech-in-noise task again. Two weeks later, participants returned for a follow-up. Effect of session (pre- or post-training) on speech intelligibility was significant, with no significant effect of group (trained, or pseudo-trained) or stimulus (audio-only, or audio-tactile). Before training, there was a significant increase in speech-envelope tracking accuracy with audio-tactile stimuli, suggesting enhanced neural representation of audio-tactile stimuli. However, there was no benefit of congruent training to audio-tactile tracking. There was an enhancement of audio-only speech-envelope tracking following incongruent training. This suggests that speech intelligibility and tracking of audio-tactile speech are not enhanced by short-term training. This work challenges prevailing assumptions by demonstrating increased speech-envelope tracking accuracy that was not linked to enhanced speech intelligibility. However, these findings are limited to short-term, bottom-up audio-tactile training. Alternate training approaches and further controls, such as tactile-only sentences or audio-only training groups, should be explored in future research.

The aim of this study was to test whether rapid invisible frequency tagging (RIFT)-a promising tool to directly measure covert attention allocated to parafoveal words-affects lexical processing during natural reading. RIFT involves flickering the background of the target word(s) at an imperceptible frequency (≥60 Hz), but imperceptibility does not necessarily equate to the absence of a cognitive effect. The current study examined the potential impact of the RIFT paradigm on eye movements. Forty-eight participants read 474 high-frequency and low-frequency target words that were embedded in one-line sentences and displayed randomly in three condition blocks: no tagging, invisible tagging at 60 Hz (i.e., the RIFT) and visible tagging at 30 Hz. Linear mixed-effect modelling and divergence point analysis revealed a robust frequency effect, with high-frequency words processed faster than low-frequency words, but, crucially, no significant differences across the three conditions in terms of its magnitude or onset latency, except for a miniscule effect on saccade amplitude (0.06˚). Overall fixation durations (irrespective of lexical frequency) did not differ between the no-tagging and RIFT conditions, but were significantly longer in the visible tagging condition, suggesting that deliberately drawing attention to the tagging area actually slowed down reading. Furthermore, a postexperiment questionnaire indicated relatively low RIFT awareness. Altogether, preserved lexical frequency effects of comparable magnitude, similar divergence points and overall fixation and saccade patterns, suggest that RIFT is a valid tool for measuring attention during reading that is unlikely to interfere with word processing and highlight its potential for application in ecologically valid settings involving eye movements.

Each year, approximately 17 000 children will have a mother who is imprisoned in England and Wales. The impact of maternal vs paternal imprisonment is very different; the negative health consequences on children are likely to be much greater if a mother is imprisoned. The number of children affected by non-custodial justice involvement is likely to be far greater given the higher number of women under probation supervision. However, no official statistics or estimates are collected for this population. The purpose of the CHAMPION (children whose mothers are involved in the criminal justice system in Dorset & Hampshire, UK - developing health and social care outcome indicators) study is to develop a better understanding of the evidence of the impact of maternal contact with the CJS (Criminal Justice System) on children and to establish a set of core outcome indicators for monitoring and evaluating affected children's health and well-being. (i) A scoping review of the health impacts of maternal vs paternal imprisonment, (ii) qualitative focus groups and interviews with adults whose mother was imprisoned when they were children and children of mothers who have either been or currently are in prison or have received community order sentences. Qualitative interviews with a range of professionals working with these children, (iii) an outcomes workshop with justice-involved mothers and professionals working with the children and families of justice-involved mothers, and adult children who experienced maternal imprisonment. Ethical approval from the University of Southampton's Faculty of Medicine Review Panel (Ref: 76946) and His Majesty's Prisons and Probation Service (HMPPS) National Research Committee (Ref: 2023-135). Publishing in peer-reviewed journals, presenting at research conferences, round table events and community-based workshops. The study's peer researchers and lived experience leaders will enable the dissemination of findings beyond academic audiences and policy makers to third sector CJS organisations and individuals with lived experience.

The neural mechanisms by which the developing brain acquires higher mathematical concepts from elementary intuitions remain poorly understood. Through a large-scale longitudinal functional MRI study of children from preschool through first and second grade, we tracked how neural responses to mathematical and nonmathematical statements change in the first 2 y of formal schooling, and we used these data to evaluate several theories of developmental change. Before school, when listening to math statements, children already engaged an adult-like cortical network, with partial specialization for geometry. Over the first 2 y of school, we observed an overall increase in math-related activation, a small recruitment of additional neural territory, reduced activation for facts that get better known, and a small overall increase in the dimensionality of representational space. fMRI responses to individual sentences suggest that these mechanisms, particularly in left inferior frontal gyrus and bilateral intraparietal sulcus, all contribute to children's growing mastery of mathematical concepts.

Neural activity has been shown to track hierarchical linguistic units in connected speech, and these responses are modulated by changes in speech intelligibility resulting from spectral degradation. In this study, we manipulated prior knowledge to enhance the intelligibility of physically identical speech sentences and tested whether this improvement would strengthen neural tracking responses. Cortical magnetoencephalography responses were recorded from 23 normal-hearing participants while they listened to intelligible speech followed by either the same (matched) or different (unmatched) unintelligible speech. When prior knowledge was available, cortical coherence at higher-order linguistic rates, particularly phrase and sentence rates, was enhanced relative to the unmatched condition and was predominantly lateralized to the left hemisphere. In contrast, cortical coherence to word-level units, which aligned with acoustic onsets, was bilateral and did not show a significant modulation by contextual information. No such coherence enhancement was observed when unintelligible speech preceded intelligible speech. This dissociation suggests that cerebral tracking of linguistic information is directly influenced by intelligibility, which itself is strongly shaped by physical speech cues. These findings provide an objective and sensitive neural index of speech intelligibility and help explain why previous studies have reported no effect of prior knowledge on cortical entrainment.

This study proposes and evaluates a two-stage large language model (LLM)-based pipeline for automated citation quality scoring in academic manuscripts. The pipeline operates as follows: in Stage 1, citation sentences are extracted from full-text PDFs and matched to their referenced articles using the Gemini 2.5 Flash model; in Stage 2, each citation-reference pair is scored for semantic relevance on a continuous 0-10 scale by a second LLM inference call operating under a structured five-tier rubric and a skeptical reviewer prompt persona. The pipeline was applied to a corpus of 121 Web of Science (WOS)-indexed engineering articles drawn from journals spanning all four Journal Citation Reports quartile strata (Q1-Q4), yielding 5,615 scored citation-reference pairs. Descriptive analysis revealed an overall mean relevance score of 7.76 (SD&#x2009;=&#x2009;2.36), with 74.7% of citations rated as Strong or Excellent. A Kruskal-Wallis test confirmed statistically significant score differences across quartile groups (H(3)&#x2009;=&#x2009;157.10, p&#x2009;<&#x2009;0.001), though the overall effect size was small (&#x3b5;&#xb2; = 0.028). Post-hoc Mann-Whitney U tests with Bonferroni correction identified Q2 articles as recording the highest mean scores (M&#x2009;=&#x2009;8.04), significantly outperforming Q1 (M&#x2009;=&#x2009;7.52), Q3 (M&#x2009;=&#x2009;7.73), and Q4 (M&#x2009;=&#x2009;7.74). The Q3 versus Q4 comparison was the sole non-significant pairing (p&#x2009;=&#x2009;0.756), indicating these strata are statistically indistinguishable in citation quality. Spearman correlation yielded a weak negative rank correlation (&#x3c1; = -0.105, p&#x2009;<&#x2009;0.001), with Q1 recording the highest proportion of Irrelevant citations (10.7%). These findings challenge the assumption that citation quality improves monotonically with journal prestige. The lower mean score of Q1 coexists with one of the highest proportions of highly relevant citations, indicating a bimodal rather than uniformly weaker profile, and a systematic annotation showed that context-dependent pointer citations are disproportionately concentrated in the Q1 Irrelevant set. We therefore attribute Q1's pattern to the broader interdisciplinary scope of top-tier articles together with a measurement effect, rather than to any single cause such as AI-assisted writing. The proposed pipeline offers a scalable, content-aware complement to existing academic integrity tools, with practical applications in editorial pre-screening and automated peer review support. An inter-rater reliability study on a stratified subsample of 150 citation-reference pairs showed strong ordinal agreement between the LLM and expert majority vote (Spearman &#x3c1;&#x2009;=&#x2009;0.643, p&#x2009;<&#x2009;0.001), with exact-category agreement of 48.0% rising to 77.3% under &#xb1;&#x2009;1 adjacent-category tolerance, and highest agreement at the Irrelevant (80.0%) and Excellent (71.0%) poles.

Adults who stutter (AWS) show a maladaptive attentional focus when speaking under socially stressful conditions and at the same time, attentional focus has shown to impact articulatory control in this population. The purpose of this study was to determine the predictive factors of social anxiety and attentional focus on articulatory control in AWS compared to a group of adults who do not stutter (ANS), varying in social anxiety levels. Seventeen AWS and 23 ANS repeated sentences in a virtual reality environment under a low- and high- socially stressful condition. Attentional focus was unsolicited and allowed to occur naturally and spontaneously. Social stress was measured using skin conductance levels changes (SCLdiff) and an anxiety questionnaire (Anxiety-Q). Attentional focus was measured using the Focus of Attention Questionnaire (FAQ). Speech motor control was assessed by measuring movement duration (DUR) and lip aperture variability using the spatiotemporal index (LA STI). Multiple regression analyses were performed to examine the effects of social anxiety (Anxiety-Q, SCLdiff) and attentional focus (FAQ) on speech motor control (DUR, LA STI). Preliminary testing showed significant differences in DUR and LA STI in the AWS compared to ANS. Multiple regression analyses showed increases in anxiety significantly predicted increases in articulatory variability but not DUR for the AWS. No significant associations were found between anxiety and attentional focus, nor for the duration of movement. Results showed social anxiety predicts changes in articulatory variability in AWS when speaking under social stress, while attentional focus plays a lesser, insignificant role.

In sentences such as "John remembered the boy took some time to rest", the locally ambiguous noun phrase (NP) "the boy" is initially parsed as the direct object of the matrix verb "remembered" (the object analysis). When the embedded verb "took" is encountered, the NP is revised as the subject of the embedded clause (the subject analysis). Open questions are how the real-time resolution of this complement ambiguity is influenced by semantic/categorial constraints (i.e., whether the locally ambiguous NP is a semantically/categorially appropriate object of the matrix verb) and selectional frequency (i.e., the frequency with which the matrix verb takes a direct object NP). The present study addressed these questions and also examined whether temporal adjuncts, which can bias parsing towards the object analysis (e.g., "John remembered the boy after&#x2026;"), influence real-time ambiguity resolution. The results showed that semantic/categorial constraints and selectional frequency drive the processor towards the subject analysis before embedded-verb disambiguation. However, temporal adjuncts gradually increased in influence and ultimately overrode these biases during processing. The observed parsing process was also simulated within an interactive constraint-based framework. Together, the experimental and simulation results suggest that real-time sentence processing is dynamically shaped by multiple competing biases.

Semantic mismatch processing is fundamental to language comprehension, yet how contextual constraint modulates this processing in Chinese reading remains unclear. The present study investigated the cognitive processing of semantic mismatch in Chinese modifier-noun constructions and examined the moderating effect of contextual constraint. Forty-eight native Mandarin Chinese speakers read sentences containing semantically matched or mismatched modifier-noun constructions embedded in high- or low-constraint contexts while their eye movements were recorded using an EyeLink 1000 Plus eye tracker. Linear mixed-effects models were used to analyze eye movement measures. Robust semantic mismatch effects were observed: mismatched constructions elicited longer first fixation durations and gaze durations, higher regression probabilities, and longer total reading times. Critically, significant interactions between semantic relation and contextual constraint emerged on gaze duration and late processing measures, with mismatch effects approximately twice as large in high-constraint contexts. Simple effects analyses revealed that contextual constraint selectively affected mismatched constructions while leaving matched constructions unaffected. These findings support predictive processing accounts of language comprehension, demonstrating that readers actively generate contextual expectations and experience amplified processing difficulty when these expectations are violated.

With growing reliance on digital health tools, the readability of online medical content is increasingly vital in patient-centered care. Artificial intelligence platforms like ChatGPT are widely used to access medical information, yet their suitability for conveying complex conditions such as multiple endocrine neoplasia (MEN) syndromes remains underexplored. This study aimed to compare the readability of MEN-related educational materials generated by ChatGPT with those from the evidence-based platform UpToDate (UTD), evaluating their appropriateness for patient education. Six related subjects that addressed MEN types 1 and 2 were chosen. The WebFX readability tool was used to examine texts produced by ChatGPT and the related UTD articles. Word count, sentence count, proportion of difficult words, Flesch Reading Ease (FRE), Flesch-Kincaid Grade Level (FKGL), and Simple Measure of Gobbledygook Index were among the metrics. R (v4.3.2) and SPSS (v25) were used for statistical analysis, and the Wilcoxon signed-rank test was used. ChatGPT responses were significantly shorter (median word count: 594.0 vs. 1955.5) and contained fewer sentences (66.0 vs. 166.0), though sentence complexity differences were not significant. Both sources scored poorly in readability: ChatGPT had an FRE of 21.3 and FKGL of 15.3, while UTD scored 24.9 and 16.7, respectively. Although ChatGPT used fewer difficult words, it had a higher proportion of complex terms (35.6% vs. 30.1%; P = 0.0306). Although both sources went over the suggested readability limits for patient materials, ChatGPT produced more condensed text. To improve understanding and promote health fairness, medical terminology must be made simpler. R&#xe9;sum&#xe9; Contexte:Face &#xe0; la d&#xe9;pendance croissante aux outils de sant&#xe9; num&#xe9;riques, la lisibilit&#xe9; des contenus m&#xe9;dicaux en ligne rev&#xea;t une importance capitale dans les soins centr&#xe9;s sur le patient. Les plateformes d&#x2019;intelligence artificielle telles que ChatGPT sont largement utilis&#xe9;es pour acc&#xe9;der &#xe0; des informations m&#xe9;dicales, mais leur aptitude &#xe0; transmettre des informations sur des pathologies complexes comme les syndromes de n&#xe9;oplasies endocriniennes multiples (NEM) demeure insuffisamment explor&#xe9;e.Objectifs:Cette &#xe9;tude visait &#xe0; comparer la lisibilit&#xe9; des supports &#xe9;ducatifs relatifs aux NEM g&#xe9;n&#xe9;r&#xe9;s par ChatGPT avec ceux issus de la plateforme &#xe0; donn&#xe9;es probantes UpToDate (UTD), afin d&#x2019;&#xe9;valuer leur ad&#xe9;quation &#xe0; l&#x2019;&#xe9;ducation des patients.Mat&#xe9;riels et m&#xe9;thodes:Six sujets portant sur les NEM de types 1 et 2 ont &#xe9;t&#xe9; s&#xe9;lectionn&#xe9;s. L&#x2019;outil de lisibilit&#xe9; WebFX a &#xe9;t&#xe9; utilis&#xe9; pour analyser les textes produits par ChatGPT et les articles UTD correspondants. Les param&#xe8;tres &#xe9;tudi&#xe9;s comprenaient le nombre de mots, le nombre de phrases, la proportion de mots difficiles, le score de facilit&#xe9; de lecture de Flesch (FRE), le niveau scolaire de Flesch&#x2013;Kincaid (FKGL) et l&#x2019;indice SMOG (Simple Measure of Gobbledygook). Les analyses statistiques ont &#xe9;t&#xe9; r&#xe9;alis&#xe9;es avec R (v4.3.2) et SPSS (v25), en recourant au test des rangs sign&#xe9;s de Wilcoxon.R&#xe9;sultats:Les r&#xe9;ponses de ChatGPT &#xe9;taient significativement plus courtes (nombre m&#xe9;dian de mots: 594,0 contre 1955,5) et comportaient moins de phrases (66,0 contre 166,0), bien que les diff&#xe9;rences de complexit&#xe9; syntaxique ne soient pas significatives. Les deux sources pr&#xe9;sentaient de faibles scores de lisibilit&#xe9; : ChatGPT obtenait un FRE de 21,3 et un FKGL de 15,3, tandis qu&#x2019;UTD affichait respectivement 24,9 et 16,7. Bien que ChatGPT ait utilis&#xe9; moins de mots difficiles en valeur absolue, il pr&#xe9;sentait une proportion plus &#xe9;lev&#xe9;e de termes complexes (35,6 % contre 30,1 %; P = 0,0306).Conclusions:Bien que les deux sources d&#xe9;passent les seuils de lisibilit&#xe9; recommand&#xe9;s pour les documents destin&#xe9;s aux patients, ChatGPT produit des textes plus condens&#xe9;s. Une simplification de la terminologie m&#xe9;dicale s&#x2019;impose afin d&#x2019;am&#xe9;liorer la compr&#xe9;hension et de favoriser l&#x2019;&#xe9;quit&#xe9; en sant&#xe9;.

The Brazilian Butt Lift (BBL) has become increasingly popular yet remains one of the highest-risk aesthetic procedures. As patients frequently use online sources to understand medical procedures, clear and reliable web-based information is essential. This study evaluates the readability, understandability, quality, and visibility of online BBL resources. A Google search for "Brazilian Butt Lift Procedure" was conducted using Startpage to reduce bias. The first 10 eligible English-language websites were analyzed. Extracted text was evaluated using 6 readability indices. Understandability was assessed with the Patient Educational Material Assessment Tool (PEMAT), overall quality with the JAMA (Journal of the American Medical Association) Benchmark Criteria. Website visibility was measured using SpyFu estimates of monthly organic traffic. Spearman's correlation was used to explore the relationship between readability and traffic. Readability for all websites exceeded the recommended sixth-grade level (mean readability score 21.68), indicating complex content. Sixty percentage of the websites scored 83% in PEMAT due to long sentences and limited explanation of technical terms; the remaining 40% websites met all understandability criteria. Journal of the American Medical Association assessment revealed deficiencies in transparency, with few sites providing references, author credentials, or publication dates. The correlation between readability and organic traffic was weakly positive (&#x3c1; = .176), indicating that frequently visited websites were not necessarily easier to read. In this review, online BBL information was generally difficult to understand and often lacks essential quality indicators. The weak relationship between readability and traffic suggests that popular websites may not provide accessible content. Improving linguistic clarity, structure, and transparency is needed to ensure patients receive reliable, comprehensible guidance on this high-risk procedure. For image description, please refer to the figure legend and surrounding text.

Typographic formats influence reading efficiency; however, knowledge remains limited regarding how these effects change across the lifespan, especially for orthographic distortions in digital environments. This study examines how conventional formats (lowercase and uppercase) and unconventional formats (mixed-case and LEET) affect reading times and the integration of meaning while reading five-word phrases. Three hundred and three adults (18-84&#xa0;years) read short sentences (five words) presented in the four formats, while reading times and memory accuracy were recorded. The results showed a graded cost pattern: conventional formats yielded the fastest reading times, mixed-case imposed moderate costs, and LEET produced the greatest slowdown and a slight reduction in accuracy. Moreover, a significant interaction between format and age was observed: although reading slowed with age in all formats, this effect was especially pronounced for LEET. These findings suggest that extreme orthographic distortions increase perceptual and pre-lexical demands, revealing limits in reading adaptation associated with aging.

While bilingual studies have extensively explored crosslinguistic transfer across various domains, prosody remains relatively underexplored. This study addresses the gap by contributing experimental data on prosodic transfer at lexical and pragmatic levels in heritage speakers of Turkish in Germany. Given typological differences between Turkish and German prosody, this population is ideal for examining whether the heritage language (Turkish) is influenced by the majority language (German) in processing prosodic information across linguistic levels. In a recent study, Zora et al. examined how majority language speakers of Turkish interpret prosodic cues, specifically lexical stress and prosodic focus, when assessing sentence acceptability. Forty participants rated sentences containing prosodic violations, revealing that lexical stress violations elicited lower acceptability than prosodic focus violations. This suggests that lexical-level prosody, rather than pragmatic-level prosody, plays a more central role in spoken comprehension for majority language speakers of Turkish. Building on these findings, this study applies the same paradigm to 40 heritage Turkish speakers in Germany. Results show the reverse pattern, such that prosodic focus violations were rated less acceptable than lexical stress violations. Typologically, this aligns with Germanic languages, where prosodic focus holds greater weight in interpretation under the same experimental paradigm. The results further revealed that proficiency modulated sensitivity to prosodic information, such that high-proficiency speakers of Turkish judged both stress and focus violations as incorrect, whereas low-proficiency speakers judged only focus violations as incorrect. These findings highlight the dynamic nature of prosodic transfer, potentially from German to Turkish, and the significant role of language dominance and proficiency in shaping bilingual prosodic processing.

Cancer survivors often experience complex and coexisting emotions throughout diagnosis, treatment, and posttreatment life. Emotion classification of patient narratives may help in understanding survivorship experiences; however, evidence remains limited for multidimensional classification using cancer survivor interview narratives. This study aimed to develop and evaluate natural language processing-based emotion classification models using Japanese cancer survivor interview narratives and to examine whether polarity and multidimensional emotion labels provide complementary perspectives. We analyzed verbatim transcripts from 15 cancer survivor interviews published by the Cancer Note, Nonprofit Organization. Survivor utterances were extracted, noninformative conversational elements were removed, texts were segmented at Japanese punctuation marks, and 5 consecutive sentences were grouped into 1 chunk. Two annotators labeled 1998 text chunks with 3-class sentiment polarity labels (positive, neutral, or negative) and multilabel Plutchik 8-emotion labels (joy, trust, fear, surprise, sadness, disgust, anger, and anticipation). Japanese BERT (Bidirectional Encoder Representations from Transformers) and LUKE (Language Understanding with Knowledge-based Embeddings) were fine-tuned to build a multiclass polarity classifier and a multilabel 8-emotion classifier. Performance was evaluated using precision, recall, F1-score, macroaveraged metrics, Micro-F1 for polarity, and Hamming loss for multilabel classification. For comparison, the same architectures were fine-tuned on WRIME (writers' and readers' intensities of emotion for their estimation), a Japanese social media emotion dataset, and evaluated on Cancer Note texts as a domain-transfer analysis. The 95% CIs were estimated using bootstrap resampling with 1000 iterations. Neutral was the most frequent polarity label, trust was the most frequent 8-emotion label, and anger was the least frequent emotion label. Label distributions were imbalanced, with most-to-least frequency ratios of 3.47 for polarity and 8.10 for 8-emotion labels. In the 3-class sentiment polarity task, interview-trained models outperformed WRIME-trained transfer models. Interview Text-BERT achieved the highest micro-F1 of 0.696 (95% CI 0.676-0.716), whereas Interview Text-LUKE achieved the highest macro-F1 of 0.660 (95% CI 0.639-0.682). In the 8-emotion multilabel task, Interview Text-LUKE achieved the highest macro-F1 of 0.427 (95% CI 0.398-0.453) and the lowest Hamming loss of 0.078 (95% CI 0.073-0.082). WRIME-trained transfer models showed lower performance, particularly in the 8-emotion task. Sadness and trust co-occurred most frequently, suggesting that positive and negative emotional elements may coexist in the same narratives. This exploratory study suggests the feasibility of domain-specific emotion classification for Japanese cancer survivor interview narratives. Models fine-tuned on target-domain narratives generally outperformed WRIME-trained transfer models, although the best architecture differed by task and metric. Polarity labels and Plutchik 8-emotion labels provided complementary perspectives on complex and coexisting emotions in survivorship narratives. However, performance for rare emotions remained limited, and the models should be regarded as preliminary research tools rather than clinically actionable systems. Larger, more diverse, prospectively or externally validated datasets, imbalance-aware methods, and user-centered evaluation are needed before clinical translation.

Higher-level action interpretation, such as inferring underlying intentions and predicting future actions, requires the integration of conceptual action information (e.g. "opening") with semantic knowledge about persons and objects (e.g. "my friend Anna", "pizza box"). However, how the neural systems for action and object recognition and memory interact with each other to form the basis for inferring higher-level mental states remains unclear. Here we use fMRI-based crossmodal multiple regression representational similarity analysis in human female and male participants to elucidate the processing stages from basic action and object recognition to mentalizing. We show that inferring intentions from observed actions or written sentences involves a modality-general network of lateral and medial frontoparietal and temporal brain regions associated with conceptual action and object representation and mentalizing. The representational profiles in these regions are explained by models capturing different types of conceptual information, revealing distinct but partially overlapping networks for action, object, and mental state representation. There was no strict separation of networks for action, object, and mental state representations, arguing against a sequential bottom-up hierarchy from action and object understanding pathways to the mentalizing network. Rather, left-hemispheric regions, specifically ventrolateral prefrontal, inferior parietal and anterior lateral occipitotemporal cortex, showed strong representational overlap, pointing towards a core network for making meaning of action-object structures at a conceptual level. We argue that this core network represents a distributional semantic hub between classic networks for action and object understanding and the mentalizing network.Significance Statement How does the human brain integrate information from actions, e.g., "open a pizza box", to understand the actions' underlying intentions? To do so, the brain needs to combine information from different neural networks-for action and object recognition-and pass them to the mentalizing network for inferring intentions, such as "satisfying hunger". We characterize the interplay of networks using fMRI-based crossmodal multivariate analyses and find that a left-lateralized core network in inferior frontal and parietal cortex and lateral occipitotemporal cortex represents all critical ingredients-conceptual action and object information as well as higher-level mental state representation simultaneously in an overlapping manner. This suggests that this core network is essential for semantic interpretation and functions as bridge between recognition pathways and the mentalizing system.

With age-related hearing loss on the rise and a large bilingual (BI) population in the United States, there is an increasing need for effective assessment tools for Spanish-speaking and BI older adults. Research in this area is limited. The AzBio Sentence Test (AzBio) is used to assess speech perception in challenging listening environments. A Spanish version was recently developed to address the needs of Spanish-speaking patients, and opens new opportunities for research with BI individuals to better understand speech perception processes in both the native and the second language. This preliminary study aims to explore the characteristics of speech recognition among older BI adults by examining the performance of younger and older proficient Spanish-English BIs on the AzBio in quiet and noisy conditions. Preliminary quasi-experimental group design. Twenty-one participants were divided into three groups: eight young American English monolinguals (M = 28.12&#x2009;years, range: 24-32), eight young Spanish-English BIs (M = 28.38&#x2009;years, range: 23-33), and five older Spanish-English BIs (M = 58&#x2009;years, range: 55-62). The AzBio was used to assess speech perception in quiet and noisy conditions in both English and Spanish, with two signal-to-noise ratios: 0 and -3 dB. Independent and paired sample statistical tests were conducted. In quiet conditions, older BIs outperformed younger ones on the Spanish version, whereas younger BIs performed better in English. However, for the older group, the linguistic advantage observed in quiet conditions disappeared in noisy environments, because performance differences with the younger group diminished when background noise was introduced. Young monolinguals and BIs showed no difference in performance on the English version. Aging poses challenges for speech recognition in noisy environments. For BIs, these difficulties extend to both languages, including the dominant one. The combined effects of bilingualism and aging negatively impact speech recognition in both languages. Understanding how individuals perform on the AzBio and other BI sentence recognition tools is essential for improving clinical assessment and intervention for those impacted by language background and age-related hearing changes.

The major purpose of the present systematic review is to critically evaluate the literature on the use of artificial intelligence (AI) with d/Deaf and hard of hearing (d/Dhh) students. It is also of interest to analyze any existing studies on the use of AI to improve the reading skills of these students. In light of limited available investigations, it is important to describe the development of a funded AI reading project to be proffered as a model for future research. The studies reviewed provide various examples of the use of deep learning, machine learning, and AI applications for d/Dhh individuals. These studies are not directly related to reading comprehension. One of the studies examined indicated that an AI-based sign language recognition technique can be used for word and sentence recognition in the future. In addition, it was hypothesized that this technique will help students improve their vocabulary. It was also emphasized that the system not only facilitates alphabet recognition, but also can be extended to the teaching of more complex reading units that form the basis of literacy. In the funded AI reading project model, reading texts were developed for d/Dhh students. The Flesch-Kincaid Grade Level (Turkish version) readability formula was used to determine whether the texts were suitable for the first- and second-grade levels. The model has the potential to serve larger groups of students through more comprehensive databases with its scalable structure.