CRC-W - Working Papers / Preprints
URI permanente para esta coleção:
Navegar
Entradas recentes
- Introducing circStudio, a Python package for preprocessing, analyzing and modeling actigraphy dataPublication . Marques, Daniel; Fernandes, Carina; Reis, Cátia; Barbosa-Morais, Nuno L.Actigraphy is a non-invasive and cost-effective method for monitoring behavioral rhythms under real-world conditions by collecting time-resolved measurements of locomotor activity, light exposure, and temperature. Although several open-source packages support specific aspects of actigraphy analysis, aspects such as preprocessing, metric calculation, and mathematical modeling are often distributed across separate software packages, limiting interoperability and increasing programming overhead. Here we introduce circStudio, a Python package that unifies actigraphy data processing and mathematical modeling of circadian rhythms within a single framework. Built from the pyActigraphy codebase and integrating circadian models from the Arcascope circadian package, circStudio provides flexible preprocessing tools, support for multiple actigraphy file formats through adaptor classes, standalone functions for computing commonly used actigraphy metrics, and implementations of several mathematical models of circadian rhythms. The package enables users to move efficiently from raw wearable data to physiologically interpretable circadian outputs. Ultimately, circStudio aims to facilitate reproducible workflows and to provide a flexible foundation for research applications across circadian biology, sleep science, and digital health.
- Introducing circStudio, a Python package for preprocessing, analyzing and modeling actigraphy dataPublication . Marques, Daniel; Fernandes, Carina; Reis, Cátia; Barbosa-Morais, Nuno L.Actigraphy is a non-invasive and cost-effective method for monitoring behavioral rhythms under real-world conditions by collecting time-resolved measurements of locomotor activity, light exposure, and temperature. Although several open-source packages support specific aspects of actigraphy analysis, aspects such as preprocessing, metric calculation, and mathematical modeling are often distributed across separate software packages, limiting interoperability and increasing programming overhead. Here we introduce circStudio, a Python package that unifies actigraphy data processing and mathematical modeling of circadian rhythms within a single framework. Built from the pyActigraphy codebase and integrating circadian models from the Arcascope circadian package, circStudio provides flexible preprocessing tools, support for multiple actigraphy file formats through adaptor classes, standalone functions for computing commonly used actigraphy metrics, and implementations of several mathematical models of circadian rhythms. The package enables users to move efficiently from raw wearable data to physiologically interpretable circadian outputs. Ultimately, circStudio aims to facilitate reproducible workflows and to provide a flexible foundation for research applications across circadian biology, sleep science, and digital health.
- Designing AI-enabled career coaching tools for women: an expert interview studyPublication . Fonolla, Sara Portell; Gaspar, Augusta D.; Correia, Luís; Pinto, Joana CarneiroAI-enabled coaching is increasingly being introduced into career development, yet limited attention has been paid to how such tools should be designed for women’s career development, particularly in male-dominated contexts. This study explored which AI-enabled career coaching use cases experts consider relevant and feasible for women, and what contextual, design, risk, and governance requirements should guide their responsible development. Sixteen multidisciplinary experts from career development, coaching, organisational development, women’s health, AI coaching, AI ethics, and DEI participated in semi-structured interviews. Data were analysed using reflexive thematic analysis. Five themes were generated: integrated expert lenses and theoretical foundations; the biopsychosocial ecosystem of women’s careers; AI architecture and mechanics of intervention; the human-AI coaching dialectic; and ethical governance and agency risk. Experts supported AI-enabled coaching only conditionally, especially for bounded functions such as structured reflection, scoped psychoeducation, behavioural rehearsal, and action support. They cautioned against tools that reproduce androcentric career assumptions, individualise structural barriers, or substitute for human empathy, contextual judgement, and professional accountability. The study develops an expert-derived framework that sets out how AI-enabled career coaching tools for women should be designed, governed, and evaluated. The framework specifies requirements relating to conceptual alignment, ecosystem attunement, intervention mechanics, and agency-centred governance. It may serve both as a guide for future tool development and as an evaluative lens for assessing existing AI coaching and career-support tools. Future research should test these bounded use cases with women end-users across different career stages and work contexts.
- Designing AI-enabled career coaching tools for women: an expert interview studyPublication . Portell-Fonolla, Sara; Gaspar, Augusta D.; Correia, Luís; Pinto, Joana CarneiroAI-enabled coaching is increasingly used in career development, yet limited attention has been paid to how such tools should be designed for women, particularly in male-dominated contexts. This study examines which AI-enabled career coaching use cases experts consider appropriate and what design and governance conditions should guide their responsible development. Sixteen multidisciplinary experts across career development, coaching, organisational development, women’s health, AI coaching, AI ethics, and DEI participated in semi-structured interviews. Data were analysed using reflexive thematic analysis. Five themes were identified: integrated expert lenses; the biopsychosocial context of women’s careers; AI intervention mechanics; the human–AI coaching relationship; and governance and agency risks. Experts supported AI use conditionally, primarily for bounded functions such as structured reflection, scoped psychoeducation, and behavioural support. They cautioned against reproducing androcentric assumptions, individualising structural barriers, and substituting for human empathy, contextual judgement, or professional accountability. The study proposes an agency-centred framework specifying requirements for conceptual alignment, ecosystem attunement, intervention design, and governance. These findings suggest that AI-enabled career coaching is most credible as a constrained, developmental support tool. Future research should examine these use cases with women end-users across career stages and contexts.
- Artificial intelligence applications supporting women’s career development: a scoping reviewPublication . Fonolla, Sara Portell; Fassi, Yasmina El; Gaspar, Augusta D.; Correia, Luís; Pinto, Joana CarneiroArtificial intelligence (AI) is increasingly integrated into career guidance and organisational decision systems, yet empirical evidence on applications designed to support women’s career development remains limited. Following PRISMA-ScR and a preregistered protocol, we searched 7 databases (plus backward and forward citation searching) and synthesised 13 empirical studies published between 2018 and 2025. Using inductive thematic analysis, we identified three functional domains: (a) bias mitigation and representation (e.g., auditing gendered language and platform-level disparities), (b) skills development and empowerment (e.g., AI-supported learning and writing interventions), and (c) career pathways and retention (e.g., matching and attrition-risk modelling). The evidence base was concentrated in system-facing applications that detect or shape inequities within recruitment, evaluation, and exposure systems; fewer studies evaluated individual-facing developmental support, and sustained career outcomes were rarely measured. Formal theory use was limited, with only a small minority of studies explicitly drawing on established frameworks; reporting on ethics, transparency, and governance was inconsistent. We suggest that research prioritises longitudinal and theory-informed evaluations, including intersectionality-informed analyses, and assess downstream impacts on women’s career trajectories alongside robust governance and accountability practices.
- Artificial intelligence applications supporting women’s career development: a scoping reviewPublication . Portell, Sara; Fassi, Yasmina El; Gaspar, Augusta D.; Correia, Luís; Pinto, Joana CarneiroBackground Artificial intelligence (AI) technologies are increasingly employed in career development interventions. However, the extent to which these technologies support women’s career advancement, mitigate structural barriers, and promote gender equity remains underexplored. Objective This scoping review systematically maps the empirical evidence on AI-driven interventions designed to support women’s career development. Methods The review was conducted in accordance with PRISMA-ScR guidelines and pre-registered on the Open Science Framework (OSF). Twelve empirical studies published between 2018 and 2025 were identified through systematic searches across seven databases. Thematic analysis was employed to inductively synthesise the data, leading to the identification of three overarching domains: (1) bias mitigation and representation, (2) skills’ development and empowerment, and (3) career pathways and retention. Results Findings suggest that AI interventions hold significant promise to reduce gender biases in career processes, enhancing women’s professional skills and agency, and facilitating sustainable career trajectories. Nonetheless, the review reveals notable gaps, including limited longitudinal data, underrepresentation of diverse populations, insufficient theoretical integration, and ethical considerations inadequately addressed. Conclusions While AI-driven interventions align with conceptual propositions about their transformative potential, empirical evidence remains fragmented. Future research should prioritise longitudinal, intersectional, and theory-informed studies to ensure AI technologies effectively promote gender equity in career development.
- Long COVID recovery and exercise adherence: 32-month studyPublication . Rolo-Duarte, Ana; Prada, Daniela; Carvalho, Ana S. M.; Borges, Ana; Bettencourt, Paulo J. G.Objective To evaluate symptom progression in COVID-19 survivors, adherence to prescribed exercise therapy, and its association with pre-infection physical activity at 21 days (T0), 6 months (T1), and 32 months (T2) post-discharge. Design Retrospective longitudinal study in a hospital-based rehabilitation unit in Portugal. The cohort included 276 patients (mean age 56.6 ± 13.5 years) with confirmed SARS-CoV-2 infection. Results Adherence was higher among patients reporting prior physical activity (48.8%; p = .003). Symptom prevalence declined over time: dyspnea (T0 = 22.4%, T2 = 7.3%), fatigue (T0 = 32.4%, T2 = 14.5%), and pain (T0 = 17.6%, T2 = 4.8%). Asymptomatic cases increased from 27.4% (T0) to 54.5% (T2). Early adherence, particularly by day 15, was associated with continued participation at day 21, and adherence at day 21 correlated with reduced dyspnea at follow-up (p = .02). Importantly, patients who remained symptomatic at day 21 took significantly longer to recover (t = –6.386; p < .001), indicating this time point as a prognostic marker of delayed resolution of exercise-modifiable symptoms. Conclusion Early initiation of individualized, structured exercise proved safe, adaptable, and associated with reduced symptom burden, especially dyspnea. Persistence of symptoms at day 21 highlights the prognostic value of early follow-up and underscores the decisive role of timely rehabilitation engagement. Structured home-based and tele-rehabilitation programs supported adherence and accessibility, reinforcing exercise as a cornerstone of long COVID management and potentially applicable to other post-respiratory rehabilitation contexts.
- Testing methodological and population-specific influences on the detection of Zipf’s law of brevity in chimpanzee gesturesPublication . Safryghin, A.; Badihi, G.; Ferrer-i-Cancho, R.; Grund, C.; Hayashi, M.; Mielke, A.; Mine, J.; Rodrigues, E. D.; Soldati, A.; Zuberbühler, K.; Hobaiter, C.; Zulberti, C.Zipf’s law of brevity is a widespread manifestation of information compression found across human languages and other communication systems. Chimpanzee gesture represents a rare absence of its expression in short-range communication. But, whether this absence reflects a feature of ape gesture production, or results from methodological obstacles or population differences is unclear. Selecting the appropriate unit of analysis is crucial for detecting linguistic patterns in any system. We assess gestural repertoires of three chimpanzee communities for Zipf’s law, while discriminating and testing different levels of unit durations (segmentation) and how finely gestural units are split (granularity). We report the first repertoire-wide detection of Zipf’s law in ape gesture in one chimpanzee community at a specific level of segmentation and granularity. We suggest that both methodological and socio-ecological factors can shape the detection and expression of Zipf’s law, emphasizing the importance of species-relevant units and metrics for meaningful cross-species comparisons.
- MindRegulation: randomized controlled trial of the effects of a relaxation and guided imagery intervention on the psychophysiological well-being, socioemotional regulation, cognitive and academic development of children in schoolPublication . Galinha, Iolanda Costa; Carvalho, Joana Sampaio; Oliveira, Ana Cristina; Arriaga, Patrícia; Gaspar, Augusta D.; Ortega, VitóriaBackground: Mental imagery has long been used in psychological therapies, but only recently did research begin to provide a scientific background for it.Imagery interventions are inexpensive anda substantial body of research supports their effectiveness on behavior change, promotion of adaptive health outcomes, anxiety reduction, and adherence to medical interventions, in both adults and children. However, literature on relaxation and guided imagery interventions benefits for children in elementary school context is very scarce. This Randomized Controlled Trial (RCT) aims to contribute to that knowledge by implementing and testing the benefits of an intervention program MindRegulation (MR) comprising relaxation, instructions for body posture, and guided imagery with socioemotional learning (SEL), conveying adaptive beliefs about oneself, the relationships with others and the environment. Method: The MR intervention will be developed in the classroom for 15 minutes before learning activities, three times per week, for five months, and its effects will be measured on a range of emotional, physiological, and cognitive outcomes. Fifteeen classes will be randomly assigned to three conditions: (a) relaxation and guided imagery-MR; (b) relaxation only; and (c) waitlist control. The RCT includes four data collection times: pretest, intermediate, posttest, and a six-month follow-up (trial registration NCT06101225, 05th October, 2023). The sample comprises 240 students, elementary school third and fourth graders, 8–11 years old. The variables measured in all times, except the intermediate, are: well-being, affect, anxiety, emotional regulation, socioemotional competencies, attention and processing speed, and perceived benefits of the intervention. Physiological indicators of emotional arousal, emotional regulation, stress and well-being are also taken, specifically, heart rate variability, electrodermal activity, actigraphy and salivary cortisol. The validity of the measures will be tested for the population and objectives of the study. Discussion: Significant improvements on the children's well-being, socioemotional regulation, cognitive function, physiological activity and academic performance are expected - after 5-months’ intervention at posttest and11 months’ follow-up -at MR condition, compared to the relaxation and control conditions. Changes in physiological activity are expected during MR and relaxation sessions. Emotional regulation, well-being and anxiety are expected to mediate the effects of the interventions over socioemotional competence, cognitive function and academic performance. Well-being and anxiety levels at pretest are expected to moderate the interventions’ effects.