6949 research outputs found
Sort by
Automatic screening of dementia and age-associated cognitive decline from conversational speech
No full textAlzheimer’s disease (AD) and Age-Associated Cognitive Decline (AACD) pose significant challenges to aging populations worldwide, making early detection crucial for timely intervention. Traditional assessments are effective but resource-intensive, limiting their scalability for widespread and casual screening. Speech and language have emerged as promising biomarkers for cognitive decline, as they reflect complex cognitive processes and can capture subtle changes in the early stage of cognitive deficits.
This thesis investigates speech and language-based cognitive screening, addressing critical limitations in existing research, including the reliance on spontaneous speech elicited through structured tasks that introduce cognitive burden, the scarcity of longitudinal studies crucial for early intervention, the predominant use of small, balanced datasets that fail to reflect real-world diagnostic distributions and the lack of deployable tools for widespread and casual testing.
Leveraging the longitudinal biographical interviews from the ILSE dataset, this work investigates both state screening (detecting current cognitive decline) and predictive screening (forecasting cognitive decline up to 12 years in advance), which is crucial for early intervention. This work introduces a comprehensive framework for processing and analyzing longitudinal, naturalistic datasets, including methods for addressing data imbalance, exploring feature selection techniques, evaluating classifier performance, and enhancing model interpretability through feature importance analysis. These methodologies provide a foundation for future research on longitudinal and naturalistic datasets.
Another key contribution of this work is the development of DemVis, an end-to-end system that integrates automated speech processing, feature extraction, and classification models to detect AD and AACD based on natural, conversational speech. DemVis offers an accessible and scalable screening tool that supports clinicians in diagnosis while also enabling casual self-assessment of both current cognitive status and future risk of cognitive decline.
Additionally, a robust multi-processing data pipeline is developed to efficiently manage large-scale speech data for speech and language-based analysis. Its modular components can be reused or adapted for future research on longitudinal and naturalistic datasets
Book of Abstracts : 3. Konferenz der Norddeutschen Wärmeforschung : 10.-11. September 2025 Hochschule Bremen
No full textDas Book of Abstracts enthält die eingereichten Kurzfassungen, welche die Grundlage für die Vorträge der Konferenz bilden
Perceptions and practices of migration and return in the Casamance, Senegal : Study Report
No full textThe qualitative research data available at RDC Qualiservcie result from research conducted in Senegal, more precisely in the region of the Casamance. The study focused on migration perceptions and practices of potential migrants and returnees and was conducted within the framework of the DeZIM research community joint project Transnational Perspectives on Migration and Integration (TRANSMIT). The research interest focused on the interactions between policy efforts to regulate migration as practices of immobilization and transnational migration processes, return and reintegration processes, and on narratives and images of migration and how those influence migration decisions and processes. In form of a multi-sited panel study, longitudinal empirical data was collected through qualitative, semi-structured interviews and one focus group interview with return migrants and potential migrants in all three departments of the Casamance during two periods of fieldwork in 2019 and 2022/23. The empirical data is available in the form of transcripts (n=53) of the audio recordings.Bundesministerium für Familie, Senioren, Frauen und Jugend (BMFSFJ
Multi-scale analysis of mass transport within regular and irregular monolithic catalyst supports
No full textStrukturierte Reaktoren, die mit monolithischen Katalysatorträgern ausgestattet sind, werden häufig für verschiedene Reaktionssysteme verwendet. Es wurde angenommen, dass die geometrischen Parameter der verwendeten Monolithen eine wichtige Rolle bei der Steuerung der Transport- und Transfereigenschaften der strukturierten Reaktoren spielen. Die unregelmäßigen monolithischen Katalysatorträger, wie z.B. offenzellige Schäume, haben komplexe und vernetzte Geometrien, die die Herausforderungen bei der Analyse der Transportphänomene in ihnen verstärken. Außerdem erfordert die Durchführung einer realistischen Analyse die Entwicklung von experimentellen und numerischen Werkzeugen, die in der Lage sind, die strukturierten Reaktoren, die mit solch komplexen und undurchsichtigen Monolithen ausgestattet sind, vollflächig zu analysieren. Die numerische Strömungssimulation (Computational Fluid Dynamics, CFD) ist ein bekanntes numerisches Werkzeug für die Untersuchung von Transportphänomenen, das im Vergleich zu experimentellen Methoden wesentlich kostengünstiger ist. Die Ergebnisse von CFD-Simulationen müssen jedoch mit den experimentellen Methoden abgeglichen werden, um ihre Plausibilität zu beweisen. Aus diesem Grund sollte ein experimentelles Werkzeug eingesetzt werden, das in der Lage ist, die oben erwähnte Vollfeldanalyse in einem solch undurchsichtigen und komplex strukturierten Reaktor durchzuführen. Da Transportphänomene in monolithischen Strukturen in hohem Maße von der lokalen Geometrie abhängen, sollte die eingesetzte experimentelle Methode nicht-invasiv sein. Unter den möglichen experimentellen Methoden ist die kernmagnetische Resonanztechnik (NMR) in der Lage, eine nicht-invasive Ganzfeldanalyse innerhalb undurchsichtiger und komplexer Strukturen durchzuführen.
In dieser Dissertation wurde der Massentransport in Festbettreaktoren untersucht, indem die Gasströmung in katalytischen Monolithen sowohl auf makroskopischer als auch auf mikroskopischer Ebene untersucht wurde. Die identische Struktur der Monolithen wurde in Simulationen und Messungen verwendet. In einem ersten Ansatz wurde die realistische Struktur der Monolithen in CFD-Simulationen implementiert, indem die Monolithen mit der μCT-Bildgebungstechnik gescannt und ihre Strukturen mit CAD-Software rekonstruiert wurden. In einem zweiten Ansatz wurde die CAD-Software verwendet, um die Monolithenstrukturen zu entwerfen und zu optimieren, die dann mit der 3D-Drucktechnologie gedruckt und für experimentelle Messungen verwendet wurden. Es wurde eine makroskopische Analyse auf Porenebene durchgeführt, um Geschwindigkeitsfelder und -profile innerhalb der Monolithen zu erhalten. Zu diesem Zweck wurden CFD-Simulationen mit der auf finiten Volumen basierenden Software OpenFOAM durchgeführt. Der Standard Navier-Stokes Solver für inkompressible Flüssigkeiten wurde für die CFD-Simulationen verwendet. Aufgrund der geometrischen Komplexität der offenzelligen Schaumstoffstruktur wurde der Abgleich der erhaltenen CFD- und MRV-Geschwindigkeitsfelder mit Hilfe einer Bildverarbeitungsmethode durchgeführt. Dieses In-house-Bildverarbeitungstool wurde implementiert, um die Qualität der Übereinstimmung zwischen den aus Simulationen und Messungen gewonnenen Geschwindigkeitsfeldern zu bestimmen. Die Wichtigkeit der numerischen Ganzfeldanalyse für den Erhalt realistischer Geschwindigkeitsfelder und -profile in offenzelligen Schäumen wurde nachgewiesen. Ein ähnliches hydrodynamisches Verhalten der Strömung im Eingangs- und Endbereich von Wabenmonolithen wurde beobachtet. Der qualitative und quantitative Vergleich der Vollfeld-CFD-Simulation und der experimentellen Ergebnisse untermauerte die Zuverlässigkeit beider Methoden für die makroskalige Analyse des Massentransports in Festbettreaktoren. Der Einfluss der strukturellen Parameter des Monolithen, der geometrischen Artefakte und der Probeflüssigkeit auf die Qualität der Übereinstimmung zwischen numerischen und experimentellen Ergebnissen wurde untersucht. Zusätzlich wurden die Ursachen für Abweichungen zwischen numerischen und experimentellen Ergebnissen und ihr Beitrag zur Diskrepanz ermittelt.
Experimentelle makroskopische Messungen wurden mit Hilfe einer NMR-basierten Methode durchgeführt, die auch als Magnetresonanz-Velocimetrie (MRV) bekannt ist. Die mikroskopische Analyse des Gasflusses wurde mit Hilfe eines räumlich aufgelösten Pulsed Field Gradient Stimulated Echo (PFG-STE) durchgeführt, um die Verdrängungsverteilung des Gases zu erhalten und den Beitrag der Diffusion zum Massentransport zu untersuchen. Diese Messungen zeigten, dass reguläre Navier-Stokes-CFD-Simulationen nicht in der Lage sind, die Diffusionsbewegung des Gases und mikroskopische Details der Monolithenstrukturen zu berücksichtigen, die in den rekonstruierten Monolithen aus μCT-Bildern nicht zu erkennen waren. Die Durchführung von Ganzfeld-CFD-Simulationen im Mikrobereich unter Berücksichtigung der molekularen Interaktion und damit der Diffusion ist für Festbettreaktoren sehr zeitaufwändig. Insbesondere bei unregelmäßigen katalytischen Monolithen wie offenzelligen Schäumen, die eine komplexe Struktur aufweisen. Es wurde die Hypothese aufgestellt, dass die Kopplung von Navier-Stokes CFD-Simulationen mit einem Partikelverfolgungslöser, der die Diffusionsbewegung von Partikeln berücksichtigt, deren Beschränkung bei der Vorhersage der Verdrängungsverteilung von Gasströmen innerhalb katalytischer Monolithen überwinden kann. Dementsprechend wurden die Ergebnisse der regulären CFD-Simulationen mit einem neuartigen Partikelverfolgungslöser disTrackFOAM gekoppelt, der am Institut für Chemische Reaktionstechnik der Friedrich-Alexander-Universität Erlangen-Nürnberg entwickelt wurde. Die gekoppelten CFD-Simulationen und disTrackFoam-Ergebnisse, die in dieser Arbeit als CFD-PT-Simulationen bezeichnet werden, zeigten im Vergleich zu den experimentellen Messungen eine gute Übereinstimmung. Das bestätigt die Zuverlässigkeit der verwendeten experimentellen und numerischen Methoden bei der mikroskaligen Analyse von Gasströmungen in Festbettreaktoren
Analyzing Urban Ecological Connectivity - A Comparative Study : Biodiverse Cities: Social-ecological Studies in Bremen-Gröpelingen II
No full text-23
Exploring catalytic reactors using computational fluid dynamics: insights and limitations
No full textCatalytic reactors are fundamental to numerous processes in the chemical industry, thus constituting the foundation for a substantial proportion of consumed products. In these reactors, reactants are converted into the desired products, while the reactions that take place are usually exothermic. Although catalytic reactors have been used extensively for over a century, our understanding of the local conditions within these reactors remains limited. However, this knowledge is essential for optimizing the processes as a whole and rendering them more economical and ecological. The field of Computational Fluid Dynamics (CFD) has emerged as a promising tool for simulating the processes in reactors with high resolution, thereby facilitating the acquisition of fundamental knowledge regarding local process conditions. Thus, the central objective of this thesis is to show the potential and the constraints of CFD in the field of reaction engineering. The work is divided into two main parts, which are built upon sequentially and differ in the type of simulations utilized.
In the first part, non-reactive CFD simulations are employed. A particular focus here is on the heat removal from the reactors, as this is an important parameter for efficient reactor operation.
The second part of this thesis focuses on the method of reactive CFD and its comparison with three-dimensional Magnetic Resonance Imaging (MRI) data. The implementation of a simplified reactor design, the integration of a novel MRI temperature measurement technique, and the utilization of ethylene hydrogenation as a model reaction allows for the quantification of three-dimensional concentration and temperature fields within the gas phase of the optically inaccessible reactor. In multi-scale CFD simulations, the reactor, including the surface-catalyzed reactions, is modeled to predict species concentrations, gas temperature, and flow field. The comparison with the MRI data demonstrates the strengths and limitations of reactive CFD, which can only qualitatively depict local conditions with high resolution
Who is Covered by Unemployment Insurance in China? De jure coverage, de facto coverage and the dynamics of development
No full textThe long period of fast economic growth in China has created entrenched patterns of dualization in the labour market, and in social security. Chinese law calls for all employment to be formalized and covered by a relatively generous system of five social insurances, including one for unemployment. However, despite stagnation in standard formal employment, the number of persons covered by unemployment insurance continued to increase in the 2010s, and the aim of this study is to explain this puzzle. The case of unemployment insurance is particularly instructive regarding dualization in the labour market, given that only employees at working age participate in the scheme. We rely on data at prefectural-city level from the National Bureau of Statistics, satellite data for night lights and urban land use, event data for labour unrest, and a corpus of regulatory documents labelled via machine learning. The analysis focuses on descriptive statistics and long-differenced regression. The results indicate that beyond classical standard employment, the expansion of unemployment insurance is driven by economic growth and cities’ competition for a shrinking pool of migrant labour.Deutsche Forschungsgemeinschaft (DFG)2
Bacterial niche differentiation in Arctic sandy surface sediments
No full textPermeable sandy sediments cover at least half of the continental margins and are regions of high biogeochemical activity. Central to remineralization processes in sediments are heterotrophic microbial communities, which possess a vast array of metabolic capabilities that allow them to use diverse substrates, ranging from complex organic matter to simple sugars and amino acids. These benthic microorganisms colonize surfaces of sand grains, as well as the interstitial porewater. This thesis aims to improve our knowledge of carbon cycling by benthic bacterial communities using a novel fractionation method which separates cells in sandy surface sediments based on their attachment to the grains. The work focuses on high latitude environments, specifically Isfjorden, Svalbard (78°N), which experiences extreme seasonality in primary productivity with prolonged periods of continuous daylight followed by a rapid transition to months of continuous darkness.
Collecting undisturbed sediment samples is the critical first step in disentangling niches and microhabitats within surface sediments. Thus, in this thesis a new small and lightweight sediment sampler, the Ellrott grab, was developed, which was designed specifically for sampling undisturbed sandy surface sediments with the porewater and overlying seawater (Chapter 2). By comparing oxygen profiles measured in situ with those in cores subsampled from the grab, I showed that the Ellrott grab causes minimal disturbance to the sample.
Furthermore, a novel fractionation method that separates microorganisms in the porewater (PW) from those loosely attached (LA) and firmly attached (FA) to the grains was developed (Chapter 3). The PW and LA cells, which comprised 3% and 8-13% of cells, respectively, were significantly enriched in aerobic heterotrophs and had faster per-cell oxygen consumption and laminarin hydrolysis rates compared to the FA fraction. In contrast, the FA fraction (84-89% of cells) was significantly enriched in anaerobes such as sulfate reducers. I hypothesize that these fractions occupy distinct niches in surface sediments: the FA fraction likely consists of cells colonizing protected areas on the grain, but are more diffusion-limited. On the other hand, the PW and LA fractions are less resource-limited but abrasion and grazing may keep their cell numbers low. These differences in composition and activity could point towards distinct contributions by these fractions to benthic carbon cycling. Next, the composition and activities of the sediment fractions across polar day and polar night in Isfjorden were compared (Chapter 4). During polar day, taxa specialized in degrading high molecular weight organic matter increased in relative abundance in the PW and LA fractions. These two fractions also showed increased laminarin hydrolysis rates. While the FA fraction remained more stable, I observed an increase in sulfate reducers in this fraction during polar night. Across seasons, Woeseia and Maribacter remained abundant in the FA fraction. I propose a partitioning of the benthic bacterial community into seasonal and stable microbial guilds, wherein the PW and LA fractions comprise the seasonally responsive communities. On the other hand, the FA fraction comprises the more stable community, potentially utilizing constantly available or less labile substrates.
The hydrolysis and uptake of mucin, an animal-derived glycoprotein, by the fractions across seasons was the focus of Chapter 5. During polar day, maximum extracellular hydrolysis rates in the PW and LA increased, while the maximum rate did not fluctuate in the FA. Selfish uptake (i.e., taking up oligosaccharides into the periplasm), was mainly mediated by Verrucomicrobiota and Planctomycetota, and was restricted to a maximum of 2% of the microbial community.
In Chapter 6, I contextualize my findings from this thesis by presenting macro-scale observations from the station in Isfjorden. Then, I synthesize the key findings presented across the chapters, and provide a cohesive overview on how this work advances our understanding of carbon cycling processes in sandy surface sediments. Finally, I propose potential directions for future research, identifying approaches that can further investigate the remaining open questions and new hypotheses generated from this thesis
Coding Legal Segmentation in Employment Law. The Worlds of Labour (WoL) Dataset
No full textThis Technical Report on the Worlds of Labour Dataset (WoL) outlines the operationalisation of data collection and coding of employment law. The Wol covers 35 indicators measuring the evolution of particular normative regulations, seven indicators measuring and defining labour law types and functions, and one that dates back to the adoption of the first labour legislation. The historical period spans from 1880 to 2022, with a focus on countries in the Global South and colonial powers. The rest of the world is covered between 1970 and 2022, or from 1991 to 2022 for former communist countries. The focus is on the three major functions of labour law: standard setting, privileging, and equalising (SPE), with partial use of the leximetric labour law database CBR-LRI (Deakin et al. 2023a).
The first part of this report describes general rules for coding the law. These Coding Instructions describe the coding objectives, principles and rules used for coding the Worlds of Labour (WoL) SPE-Index on the standard-setting (S), privileging (P) and equalizing (E) functions in worldwide national labour legislation. The second part, the Description of Indicators, contain the two template versions used, including coding instructions by variable for both WoL-SPE and CBR-LRI variables. The third part lists technical information on the countries covered by WoL, including sources for the introduction of labour regulation. This technical paper is complementary to the conceptional and theoretical articles concerning the database (Dingeldey et al. 2022; Carlino et al. 2025) and the theoretical background on legal segmentation (Mückenberger and Dingeldey 2022; Fechner et al. 2025). The commented dataset country by country as described here is published in GESIS as „WoL – The Worlds of Labour Dataset“ (Fechner and Carlino 2025).2
Contextualizing qualitative research data for secondary reuse — a guide for researchers on the preparation of a study report.
No full textThe contextualization of qualitative research data for further scientific use is one of the central challenges in data sharing (see Medjedović & Witzel, 2010; Corti et al., 2014; Medjedović, 2014). With this guidance document, the Qualiservice Research Data Center would like to support you in contextualizing your research data. On the following pages, you will find rec-ommendations for the description and documentation of your research materials. The focus is on writing a study report that describes your research project, the origin of the data, and the data itself. Section 2 explains the study report in more detail. Section 3 presents a template for writing the study report that can serve as a guide and be adapted to the requirements of your research project. The focus here is on the presentation of qualitative interview studies. Other types of qualitative research data might require more comprehensive adaptations. Qualiservice materials to provide support in presenting other studies are in preparation. As an introduction to the topic, the following section provides an overview of the importance of contextualizing the data and the process of contextualization at Qualiservice