Skin aging is spatially heterogeneous. Chronological age and most systemic aging clocks provide useful population-level estimates, but they do not explain why adjacent or anatomically distinct skin regions may show different rates of dermal thinning, matrix remodeling, pigmentation change, vascular fragility, inflammation, or treatment response. We define the Predictive Dermatome as a computationally derived, anatomically localized skin region characterized by a shared or probabilistically related biological aging state and trajectory inferred from integrated molecular, structural, imaging, clinical, and exposomic data. Unlike a conventional neurological dermatome, which denotes a cutaneous territory associated predominantly with sensory innervation from a spinal nerve root, the Predictive Dermatome is not defined by nerve distribution. The proposed framework integrates spatial transcriptomics, spatial proteomics, single-cell or single-nucleus reference maps, optical coherence tomography, confocal or ultrasound imaging, histology when available, tape-strip biomarkers, extracellular matrix and glycation-related measurements, and exposomic data streams. We describe the learning task, input-output structure, spatial graph construction, modality-specific encoders, transformer-based token fusion, training objectives, regularization strategies, and missing-modality handling. Rather than requiring spatial omics or tissue sampling for every clinical patient, the framework uses transcriptomic, proteomic, histologic, and extracellular matrix measurements as discovery and validation layers from which clinically deployable surrogate models could be trained using non-invasive and minimally invasive inputs. We also outline preprocessing steps for high-dimensional spatial transcriptomic data, cross-modal image-omics registration, validation stages, uncertainty estimation, and ethical safeguards including privacy-preserving learning and bias auditing across skin phototypes. The Predictive Dermatome should therefore be viewed as a staged research architecture for precision dermatology, not as a currently validated clinical decision system. Its value lies in defining how molecular geography, imaging, and environmental exposure may eventually converge to support localized risk stratification, longitudinal monitoring, and biologically guided intervention in skin aging and regenerative dermatology.
Sleep-disordered breathing, and obstructive sleep apnea in particular, places sustained demand on specialized units, leading to prolonged waiting lists that in some settings exceed one year for diagnostic tests such as respiratory polygraphy (PLG). In a previous study, a discrete-event simulation (DES) model was developed and validated as a digital twin of the patient flow in the Sleep-Disordered Breathing Unit of the Pulmonology Department at Hospital Álvaro Cunqueiro (Vigo, Spain), yielding a weighted Relative Percent Difference (RPD) of 3.6% for the mean number of patients in queue and 34.6% for the mean waiting time relative to real-world data. However, the transition probabilities governing patient routing across care stages were manually calibrated through a time-consuming, iterative process based on expert judgment, which limits the reproducibility and optimality of the solution. To address this limitation, the present study proposes and evaluates a simulation-based optimization (SBO) framework coupled to the DES model, implemented in MATLAB©, with the goal of automating the calibration of the state-transition matrix. In the main experiment, aimed at calibrating the mean number of patients in queue, the optimal solution reduced the global weighted RPD for this indicator from 3.6% under manual calibration to 1.11%. Complementary experiments that incorporated waiting time into the objective function or redistributed stage weights toward PLG confirmed the ability of the SBO framework to prioritize clinically relevant stages, at the potential cost of increased aggregate error in the queue indicator. To the best of our knowledge, this is the first application of SBO to the automated calibration of a state-transition matrix in a DES model of a sleep-disordered breathing unit, offering a quantitatively effective alternative to manual parameter tuning.
IntroductionDepression and other mental health challenges can affect daily functioning and well-being, while social stigma and limited access to culturally appropriate digital resources may discourage individuals from seeking support. This study presents REWAA, a bilingual mobile application designed to support emotional self-tracking, access to mental health resources, and interaction with mental health professionals in Arabic and English.MethodsREWAA was developed as an Android application with separate patient and therapist interfaces. Its features include mood self-tracking, an exploratory AI-based facial-analysis tool, the Bilingual Hybrid Mood Heuristic (BHMH) protocol, an AI-supported conversational tool, mental wellness resources, peer support groups, and therapist communication. The application was evaluated through a user survey involving 200 respondents to assess usability, perceived usefulness, emotional awareness, privacy, and user acceptance.ResultsSurvey responses indicated generally positive perceptions of REWAA. Most respondents rated the graphical user interface as comfortable and easy to navigate (70%), while 60% reported that the REWAA bot helped them better understand their feelings and provided practical advice. In addition, 55% reported a significant improvement in emotional awareness, and 65% considered the application useful alongside professional therapy. Overall, the findings indicate positive user perceptions of the application's accessibility and integrated support features.DiscussionREWAA demonstrates the potential of a bilingual digital platform to bring emotional self-tracking, supportive resources, and therapist interaction together in one application. However, the findings primarily reflect user-reported usability and acceptance and do not establish clinical diagnostic accuracy or therapeutic effectiveness. Further evaluation involving appropriate clinical validation and longitudinal follow-up is needed to assess its effectiveness and reliability in real-world mental health settings.
IntroductionThe therapist-guided, webcam-based OMPRIS treatment for internet use disorder (IUD) has demonstrated efficacy in prior studies. Research has also shown that comorbidity, motivation for change, and therapeutic alliance influence treatment outcomes in general. In a secondary analysis of the registered randomised controlled trial (RCT) of OMPRIS, we tested whether these psychological factors relate to IUD symptom severity (IUDSS) and inter-individual differences.MethodsParticipants of the OMPRIS study (N = 180, 136 men [75.6%], 41 women [22.8%], 3 others [1.7%]) Mage 31.78 years, SD = 12.19, range = [16; 84]) were tested at six measurement times (pretreatment—2 years after). Primary outcome: IUDSS. Independent variables: therapeutic alliance, depression, anxiety, and motivation for change. We used multiple imputation for missing data, mixed-effects models adjusted for sociodemographics and IUD subtype, and cluster analyses for hypothesis testing.ResultsLinear mixed-effects models showed that therapist-rated alliance (negative), depression (positive), anxiety (positive), recognition (positive), and taking steps (negative) were associated with IUDSS, whereas ambivalence, patient-rated alliance, sociodemographics, and IUD subtype were not. Exploratory clustering of individual IUDSS slopes identified three treatment-response groups (Good, Low, and High). These clusters differed in baseline severity, comorbidity, and motivation. High Responders showed the highest recognition level (strongest psychological correlate), reflecting reduced denial and greater readiness for change.DiscussionThis study identified psychological factors associated with IUDSS trajectory. Findings underscore that comorbidities can delay strong treatment success. In general, inter-individual differences in treatment response should be considered by therapists.Clinical Trial Registrationhttps://drks.de/search/en/trial/DRKS00019925/details, identifier DRKS00019925.
BackgroundGeneral-purpose large language models (LLMs) have become a popular source of oral health information. However, the patient education materials (PEMs) generated by these models often pose readability challenges. This study aimed to: (1) compare the readability of PEMs generated by six AI chatbots in response to ten frequently asked questions on pediatric oral health; and (2) determine whether the readability of these texts meets the level recommended by the American Medical Association (AMA).MethodsTen sample questions on pediatric oral health were compiled. Each question was submitted once to each of six AI chatbots (GPT-5 Mini, Gemini 3 Flash, Claude Haiku 4.5, DeepSeek-V3, Kimi K2.5, and Qwen3-Max), generating a corpus of 60 texts in a blocked (repeated-measures) design. Six readability indices — Flesch Reading Ease (FRE), Flesch–Kincaid Grade Level (FKGL), Gunning Fog Index (GFI), Coleman–Liau Index (CLI), Simple Measure of Gobbledygook (SMOG), and Automated Readability Index (ARI) — were calculated for every text. Each index was analysed separately using the Friedman test, with the ten questions as blocks and the six models as conditions, followed by Dunn's multiple comparison test. Kendall's coefficient of concordance (W) was calculated to quantify the consistency of model ranking across questions.ResultsDeepSeek-V3 showed the highest readability (easiest texts) and Gemini 3 Flash the second highest, whereas Claude Haiku 4.5 showed the lowest readability across all six indices; GPT-5 Mini, Qwen3-Max, and Kimi K2.5 occupied intermediate positions. All six Friedman tests were statistically significant (χ2(5) = 31.77–35.94, all p < 0.0001), and Kendall's W ranged from 0.635 to 0.719, indicating that the relative ranking of the six models was broadly consistent across the ten questions. Against the AMA-recommended sixth-grade reading level, all models exceeded the recommended difficulty on every index; even against a more lenient seventh-grade threshold, only the ARI median of DeepSeek-V3 (6.5) fell below 7.ConclusionsThe PEMs generated by the six LLMs were written at a reading level above the AMA recommendation, and reading difficulty differed systematically among models. Readability should be explicitly controlled through prompt design, model optimisation, and professional review before such materials are used in pediatric oral health education.
