In silico neuroscience is an emerging research paradigm consisting in the generation of large amounts of in silico neural responses using encoding models, followed by the experimentation on these in silico data. The in silico generation of neural responses is resource-efficient, saving enormous amounts of time and money, and therefore greatly speeds up research. This dissertation consists of my three doctoral studies, each contributing to the paradigm of in silico neuroscience in an idiosyncratic way, with a focus on human vision. In Study 1, I collected and released THINGS EEG2, a large-scale visual neural dataset (LSVND) of EEG responses of 10 subject to 16,740 naturalistic images. I showed that this large amount of data allows to train accurate encoding models of neural responses to images–including end-to-end encoding models based on randomly initialized deep learning architectures–indicating the suitability of this EEG dataset for in silico neuroscience. In Study 2, I used the paradigm of in silico neuroscience to develop relational neural control (RNC), a neural control algorithm to move from an atomistic understanding of visual cortical areas (i.e., “What does each area represent?”) to a network-level understanding (i.e. “What is the relationship between representations in different areas?”). Crucially, I then validated the in silico discoveries on in vivo fMRI responses recorded from independent subjects, thus revealing how visual areas jointly represent the world as an interconnected network, and showcasing the power of in silico neuroscience. In Study 3, I contributed to the improvement of the out-of-distribution (OOD) prediction accuracy of encoding models by releasing NSD-synthetic, the OOD companion dataset of the Natural Scenes Dataset (NSD)–the most popular LSVND of human fMRI responses to naturalistic images. I showed that encoding models trained on NSD generalize in-distribution (ID) and, to a lower extent, OOD on NSD-synthetic. Furthermore, these OOD tests revealed differences between encoding models that were not detected ID. Thus, NSD-synthetic provides unique insight for the development of more robust encoding models that better generalize OOD, in turn increasing the reliability of in silico neuroscience findings. Together, my vision for the future of the paradigm of in silico neuroscience is that richer LSVNDs including OOD components will promote encoding models that more accurately predict in silico neural responses, leading to an increase of in silico experimentation that will ultimately result in faster scientific development.

We develop and apply generalized Langevin equations (GLEs) for modeling non-Markovian dynamics in discrete time-series data. Emphasizing both theoretical and computational aspects, we investigate how memory effects, non-Gaussian noise, and coupling between multi-dimensional reaction coordinates shape the dynamics of stochastic observables. Through diverse examples from synthetic, simulated, and experimental data, we show that the GLE framework offers a versatile and practical tool for data-driven modeling and prediction of complex dynamics, and establish it as a physically grounded and efficient framework for time-series analysis across molecular, environmental, and financial systems.

We begin by outlining the theoretical basis of the GLE via the projection operator formalism and present numerical techniques for extracting parameters from discrete data, suitable for analyzing systems in and out of equilibrium. In the molecular dynamics setting, we apply these methods to dihedral transitions in butane using molecular dynamics (MD) trajectories to compare various GLE formulations. Incorporating non-Gaussian noise improves predictions of mean first-passage times, surpassing Gaussian-based Markovian embeddings. For multi-dimensional systems, we develop GLE models for the coupled dynamics of the dihedral angles of pentane, revealing strong off-diagonal friction from intramolecular interactions. These models capture cross-displacement statistics and reaction kinetics in MD data.

For real-world applications, convolution filtering decomposes time-series data, and we fit a GLE model to the fast-fluctuating component. This reveals long-range memory in daily weather data, while financial assets exhibit rapid decay, consistent with the efficient-market hypothesis. GLE-based forecasting rivals the performance of state-of-the-art methods like recurrent neural networks while being more computationally efficient and offering physically interpretable insights.

As part of time-series analysis, we explore the microscopic origins of memory friction $\Gamma(t)$. We begin by the computation of static friction coefficients from MD simulations and introduce an integration-based method that accurately estimates friction and separates electrostatic and Lennard-Jones contributions to diffusivity in water. Next, we examine frequency-dependent friction spectra for water and methane in bulk water. Water spectra align with hydrodynamic theory for a diffusing sphere in a viscous medium, enabling feature attribution of $\Gamma(t)$, while methane reveals deviations that can be rationalized by the concept of a viscoelastic hydration shell. We extend these findings to polymeric hydrogels, where a generalized Stokes-Einstein relation considering an interfacial adsorption shell accounts for differences between viscosities from experimental macro- and microrheology data.

A reverse genetics system (RGS) for iridovirus frog virus 3 (FV3, Ranavirus rana1), one of the key members in Megaviricetes class, was successfully developed using transformation- associated recombination (TAR) cloning in yeast Saccharomyces cerevisiae. This method enabled the construction of BAC-YAC constructs containing the full-length FV3 genome, from which infectious virus could be rescued. One of the constructed clones - FV3-S3Ar - retained key phenotypic and functional characteristics of the parental FV3, demonstrating the integrity of the construct. To evaluate the established RGS applicability, a recombinant virus was generated by replacing ORF64R with an enhanced green fluorescent protein (EGFP), constructing a reporter virus FV3-Δ64R-EGFP. The resulting reporter virus showed stable replication and persistent EGFP expression over several passages. This confirmed the platform’s suitability for targeted gene manipulation and the viability of the system. An alternative rescue method using the heterologous fish ranavirus Largemouth bass virus virus (LMBV, Ranavirus micropterus1) as a helper virus was developed and validated. This approach proved effective for virus rescue and may provide a safer and more straightforward option compared to the use of a traditional UV-irradiated helper virus. Overall, this system offers a reliable and flexible tool for the genetic mutagenesis of FV3 and paves the way for future investigations into viral gene function, pathogenicity, and potential vaccine development.

Feed management decisions play an essential role in reducing greenhouse gas (GHG) and nitrogen (N) emissions from ruminant farming systems. However, evaluating the downstream effects of diet on emissions in dairy production is challenging. This is due to the complex interplay among interconnected components such as animals, housing, manure storage, and soil. Consequently, a comprehensive assessment that considers both direct and indirect GHG and N emissions and accounts for the underlying processes and drivers of carbon (C) and nitrogen within the system is necessary. Static emission factors (EFs) and empirical models often fail to capture the spatial and temporal variability of these systems. This limits their utility for site-specific assessments and targeted mitigation strategies. This thesis addresses this problematic by applying PB modelling approaches that simulate the underlying biogeochemical processes driving emissions, thereby offering a more dynamic and comprehensive perspective. The research begins with a comprehensive review (the first Paper) of modelling the influence of feed management on GHG and N emissions in cattle farming systems. It contrasts statistical and empirical models, with mechanistic PB models. The former are useful for inventory applications, whilst the latter capture dynamic interactions between feed composition, microbial fermentation, and downstream emissions. This review establishes the scientific rationale for integrating PB approaches into whole-farm emission assessments. By identifying the limitations of simpler methods, it sets the stage for a more sophisticated modelling approach. Building on this foundation, the second paper introduces an innovative PB modelling framework. This framework creates links across the manure management chain. It integrates the Dutch Tier 3 model for animal emissions with the Manure-(DNDC) model for manure storage and a soil biogeochemistry model. This study evaluates the downstream impact of dietary factors on whole-farm emissions. The methodology was rigorously applied to two contrasting dairy systems: a confined system in Germany and a pasture-based system in New Zealand. The differences in production systems between them demonstrated significant differences in C and N emission. This highlights the critical role of feeding and manure management practices in shaping the environmental footprint. This application showed the framework's adaptability and its capacity to reveal system-specific dynamics. The third paper further advances the research by comparing the dynamic PB Tier 3 modelling approach with IPCC Tier 1 and Tier 2 methods. This comparative analysis showed substantial differences between PB Tier 3 dynamic EFs and IPCC Tier 1 and 2 EFs. This reveals that PB models can capture interannual and system-specific GHG emission variability more accurately, leading to more precise and adaptable emission estimates. The housing component of the Manure-DNDC model was validated against field measurements, confirming its effectiveness in simulating ammonia and methane barn emissions, a key source of uncertainty in GHG assessments. This validation underscores the reliability of the PB approach for capturing complex emission processes that generic EFs may overlook. Collectively, these studies demonstrate that integrating PB models into whole-farm emission assessments enables a more nuanced understanding. It shows how feeding and manure management practices impact GHG and N emissions. These models offer a more wholistic and accurate picture of on-farm emissions by simulating the underlying processes and considering the interactions between different farm components. This integrated modelling approach has significant implications. It can help develop of targeted mitigation strategies and improve the C accounting in the context of Carbon-farming and monitoring, reporting and verification (MRV) in agricultural systems. It can also inform future policies aimed at reducing the environmental footprint of cattle farming. The PB models represent a step towards more precise and adaptable tools for researchers and policymakers. They facilitate evidence-based decision-making in sustainable agriculture and pave the way for more environmentally responsible livestock production.

Helminths contribute a large proportion of the global burden of disease. Mainly by causing gastrointestinal infections of varying intensity in humans and animals. We asked whether commensal bacteria in the host gut can modulate the immune response against the gastrointestinal helminths, whether depletion of the microbiota with antibiotics would affect anthelmintic type 2 immune responses in the rodent model of hookworm infections. Our results showed that antibiotic treatment resulted in microbiota depletion, but neither worm burden nor female egg production was affected. Upon secondary infection with H. polygyrus, worm burdens increased in antibiotic-treated mice, suggesting that a lack of microbiota had created a supportive environment for the worms to establish themselves. Immune responses were not affected by microbiota depletion. Similar type 2 cytokines and Th2 cell frequencies were found during primary infection, during the memory phase, and after challenge infection. Interestingly, tissue-resident memory cells expressing CD69 found in the peritoneal cavity were significantly enriched during secondary infection, suggesting that the peritoneum is a strong reservoir for memory cells. Antibody responses similar to cellular responses did not change in the presence or absence of microbiota. The only increase in antibody isotype IgE was associated with the memory phase of the response, suggesting a possible enhancement of IgE responses due to the lack of bacteria coupled with the accumulation of allergens from the worm. The host and worm metabolic changes analyzed using NADPH fluorescence lifespan imaging also showed no significant changes in the presence or absence of microbiota. Overall, type 2 immune responses against a gastrointestinal nematode as well as metabolic changes of hosts and nematodes are robust and no effect of the presence or absence of microbiota was shown by prolonged antibiotic treatment.

On the other hand, this work investigated the different intensity of infection of gastrointestinal helminths in natural hosts by analyzing the hepato-tracheal migration of Ascaris suum in two different mouse strains. A susceptible C57BL/6 mouse and a resistant CBA mouse. We asked whether the tendency of individuals to be more or less susceptible is related to different innate immune responses elicited during tissue migration. Interestingly, contrary to our expectations, we found that the susceptible strain had high levels of inflammation, while the resistant strain had no inflammation at all. The inflammation caused by the susceptible strain was associated with the expression of type 2 cytokines, namely IL-5 and IL-13, by innate lymphoid cells, which subsequently responded to the recruitment of eosinophils and alternative (M2) macrophages to the mice lungs followed. Eosinophil-depleted mice were subsequently examined. The results showed that the presence of eosinophils was not related to protection, as reflected in similar larval burdens between eosinophil-deficient and eosinophil-sufficient mice. Because we detected high levels of eosinophil peroxidase in the serum of susceptible mice at the lung stage of infection, these cells may be more affected by the pathology. In conclusion, these works showed that a strong type 2 immune response is associated with susceptibility, while the resistance associated with CBA mice requires further investigation since no immunological parameter analyzed indicates an increased susceptibility to the infection.

The small intestine (SI) plays a crucial role in nutrient absorption, ensuring systemic energy homeostasis while serving as a protective barrier against pathogens and external insults. The SI epithelium is organized into crypt-villus units, where crypt-resident stem cells give rise to differentiated cell types along the villus. Among all differentiated cells, enterocytes are the most abundant and primarily responsible for nutrient sensing and uptake. However, the molecular mechanisms governing enterocyte differentiation and functional specialization remain poorly understood. Here, we identify the transcription factor c-Maf, a member of the AP-1 family, as a key regulator in enterocyte differentiation and functional zonation along the intestinal crypt-villus axis. We demonstrate that c-Maf is specifically expressed in small intestinal enterocytes, and exhibits a zonated expression pattern, confined to the mid-villus region. We show that this spatial restriction is regulated by the increasing BMP signaling throughout the villus and overlaps with the expression region of most nutrient transporters. To investigate the role of c-Maf in enterocyte biology, we used a constitutive intestinal epithelial cell (IEC)-specific c-Maf knockout mouse model (MafΔIEC). These mice were viable and born in normal Mendelian ratios but exhibited impaired enterocyte maturation, reduced nutrient uptake, weight loss and decreased body temperature. Transcriptomic and functional analyses revealed reduced expression of nutrient transporters involved in carbohydrate and protein absorption, as well as reduced absorptive capacity as shown by in vitro and ex vivo uptake assays. Furthermore, c-Maf-deficient enterocytes showed defects in the adaptation to increased carbohydrate availability. Besides its role in nutrient absorption, c-Maf deletion also impacted immune and microbial homeostasis, leading to a reduction of intraepithelial lymphocytes (IELs) and an expansion of segmented filamentous bacteria (SFB). Our ongoing work explores the role of c-Maf in dynamic states of intestinal epithelial and nutritional remodeling, such as during maternal lactation or chronic intestinal inflammation. Collectively, our findings establish c-Maf as a critical transcriptional regulator of enterocyte differentiation and function, essential for intestinal homeostasis, nutrient uptake and host-microbiota interactions.

Intrinsic disorder in proteins is a door to an entire section of biology that had remained invisible for many years, escaping comprehensive study. Over time, as the methods to probe disorder became better tailored to the task, it appeared that intrinsically disordered proteins (IDPs) were just as important and central to biological processes as any other known factor. The focus of this work revolves around two main axes. The first is the application of Nuclear Magnetic Resonance (NMR) to the study of the clathrin assembly protein 180 (AP180), an intrinsically disordered protein central to the phenomenon of clathrin mediated endocytosis in neurons. The second axis is about improving a conformational ensemble-building strategy for the study of disordered proteins by incorporating data from different methods, namely NMR, single-molecule Forster resonance energy transfer (smFRET), and small-angle X-ray scattering (SAXS). Clathrin-mediated endocytosis (CME) is one of the most prominent ways for eukaryotic cells to interact with their environment, this process allows them to internalize external cargo as well as membrane and membrane associated components into vesicles, or clathrin-coated vesicles (CCV) in this case. However, the molecular fine-tuning of the process escapes exhaustive description, since a number of the many actors involved bear heavily disordered regions, who are by essence challenging to study. AP180 is involved in the early stages of neuronal CME and is one of the most disordered proteins in the system. It does bind the membrane through a folded domain that has been extensively characterized, and yet, most of the knowledge about its long disordered cytosolic tail is limited to having located a number of hypothetical binding regions, i.e. short consensus sequence stretches of 3 to 15 residues, known to promote interaction with other main actors of CME. These other actors are the clathrin protein complex, the primary scaffolding unit of the CCV, and the heterotetrameric clathrin adaptor protein 2 complex (AP2). Here, the ~600 residue-long disordered AP180 was expressed in different ~200 amino-acid long and slightly overlapping constructs to facilitate protein expression as well as NMR amide backbone assignment, which hinted at a typical disordered chain without long-range self interactions. Numerous partner titrations were performed and analyzed by 2-dimensional 1H-15N correlation NMR experiments, and were conducted across all AP180 stretches as well as the full intrinsically disordered region (IDR), and under different conditions to best accommodate for the needs of the tested interaction partners. The three most tested partners were the alpha and beta appendages of AP2 (AP2α and AP2β2), and the clathrin heavy chain terminal domain (CHCTD), all being known targets of the consensus binding sequence motifs that exist along AP180. These titration experiments, analyzed through the scope of chemical shift perturbations, intensity ratios, and R1ρ relaxation, first confirmed the relevance of the motifs, since the expected partner proteins did bind to their associated target. This finding supported the idea that AP180 plays a role in early endocytosis by allowing for weak interactions to be made increasingly more relevant closer to the v membrane as the local gradient of existing motifs in space gets denser, increasing the number of binding events. Binding affinities could be estimated on a residue-wise basis by using local R1ρ relaxation rates, showing that motifs mostly shared the same binding strength, differences could be ascribed to the nature of the residues involved as well as broader sequence context, but all motif affinities seemed to be in the same order of magnitude (hundreds of micromolar). An interesting observation arose from the apparent ability of motifs to bind the other partners as well as their known target, pushing forward the premise of motif promiscuity, useful to endocytosis in the context of maintaining a network of interactions between all partners. A considerable finding was unraveling an unknown, extended and strong binding site to AP2β2, dubbed the extended interaction motif (EIM). This extended site is composed of two neighboring regions within AP180-IDR (residues 435-445 and 470-500), and shorter constructs of AP180 were designed around this specific region for further investigation. The AP180 EIM – AP2β2 experiments yielded an exchange rate of 662 ± 35 s−1, a percentage of bound population of 8.9 ± 0.97%, and an affinity toward the appendage in the order of only a few micromolar, much stronger than what was observed for the individual motifs. This affinity as well as the exchange rate were calculated for the entire binding region rather than individual residues, this was achieved by acquiring and fitting NMR relaxation dispersion data, and the affinity was also later validated by isothermal titration calorimetry, confirming that the EIM drives the overall interaction between AP180 and AP2β2. It was also shown that upon saturation of AP180 with excess AP2β2, the motifs binding events became relevant again, suggesting a context-dependent interaction pattern, where the EIM plays a larger role in earlier CME, i.e. lower-concentrated environments, where the motifs could engage in crowd-control, interacting with all partners available and tightening the network as the vesicle forms. The first hypothesis for the role of this EIM was thus the potential for AP2 recruitment or AP2 sensing at the nascent vesicle pits. This stronger binding region also happens to span the area within AP180 that is the most bereft of observed clathrin binding events, suggesting that in a context where both AP2 and the clathrin are within reach of all AP180 motifs, such as the later stages of vesicle formation, AP180 could be concertedly binding both partners without hindrance through the EIM and the motifs, tightening the grip of the vesicle on the coat components by exploiting all the numerous binding sites. While motifs could indeed predict binding events, it appeared clear through both the discovery of the EIM and the promiscuity of the motifs that predictions were not enough to assess the behavior of this IDP, supporting the importance of thorough screening of the interaction landscape, and highlighting the need for developing tools that are able to efficiently provide these insights. This study has been published, and a summary as well as the publication can be found in Chapter 5. As mentioned above, the second axis of this work focuses on the development aspect of a strategy to better study disordered proteins. Describing IDPs properly can be a tedious task, and one of the best ways to achieve this is through the use of conformational ensembles ; collections of different conformers of vi an IDP population, expected to accurately capture the properties and favored dynamics of the population of interest. Building an ensemble requires the proper computational tools, but also high quality experimental data to use as input. The different methods of NMR perform remarkably well at providing such data, ensembles generated by NMR inputs cover most of the relevant timescales for the dynamics of IDPs, but lack high-quality descriptions of the furthest-reaching distance regimes, with NMR hitting a limit at around 2.5 nanometers, which is too short to inform on structural parameters such as the overall shape and size sampled by a protein. That weakness can be nicely palliated by using restraints from other techniques in the generation of ensembles. One such technique is smFRET, with a sensitive distance regime of ~2.5 to ~10 nanometers, synergistic with NMR. Part of my project revolved around aligning, characterizing, and then using a confocal single-molecule FRET setup in order to incorporate long-distance restraints to an ensemble-building approach. The system of interest was a 110 residue-long disordered stretch of the Nterminus of the measles virus phosphoprotein, and the incorporation of longer-range distances produced more reliable output ensembles that provided the new insight of an overall higher compaction of the protein than previously thought. The setting up process is presented in Chapter 6, and the results of the proof-ofconcept have been published, they are summarized alongside the publication in Chapter 7.

Understanding what animals want is essential, as it reflects their behavioural mechanisms and needs – particularly interesting in the context of animal welfare. However, studying animal decisions is challenging because the animals cannot just be asked for their preferences. While previous studies have primarily focused on studying animals’ choices and preferences for specific resources, they have rarely examined daily decisions. This study introduces a novel testing system, the SW, which optimises and extends prior-choice tests in various ways: Choices between eight simultaneously offered ecologically relevant resources are provided in a long-term test to study everyday moment-to-moment decisions with the example species of rats (Long Evans, Rattus norvegicus). Each of the 11 rats is once used as focus rat for ten consecutive days, whereas the other 1-2 sister/s serve as social partners. The SW system consists of a CCC connected via plastic tubes to eight resource cages. These cages provide a quasi-natural environment including resources for basic survival (food, water, hiding/resting), resources for more specific needs (running wheel, sand bath) and resources of additional ecologically relevant stimuli (predator odour, novel object, restricted contact to social partners). Movement sensors are installed at the entrances and exits of the resource cages so that data of the daily frequencies, durations and sequences of visits, as well as for the decisions and choices of the various resources is collected automatically. The “choices” are defined as the duration between the time of entry to a resource cage and the entry into the next resource cage. The “decisions” refer to the duration between the exit from one and the entry into the next resource cage. The data are analysed using generalised mixed-effects models including survival modelling for multiple states. Results show that the general activity decreases from day 1 to 2 and remains on a constant level throughout the experimental phase. Rats visit all resource cages at least once per day throughout the experiment. The frequency of resource use is relatively high in the RWC, NMC, SPC; intermediate in the FIC and WIC and lowest in the POC, NOC and DAC. The daily duration of resource use remains relatively stable across the experiment but vary across the different resource types. The NMC constantly shows high duration of use, while POC, NOC and DAC have constantly low durations with peaks on days when stimuli are renewed (e.g. fresh predator odour, a new novel object, or treats in the sand bath). The most frequent transitions occur between the RWC, WIC, FIC, and SPC. The frequency and duration of the visits do not necessarily reflect the importance of the resources. As depending on the need, a short visit e.g. to the WIC, may be sufficient to satisfy the animals' need to drink. Nevertheless, the fact that all resources are visited daily suggests that each holds relevance for the animals. The study also evaluates whether the experimental system induces stress. Potential stressors include the novelty of the experimental system and resources, spatial and social restrictions, and the limited food choice. Stress is measured non-invasively via FCM measurement. The mean FCM concentration decreases slightly when the rats are transferred from the housing cage to the experimental system and slightly increases when they are returned to their housing cage. Furthermore, the rats quickly acclimatise to the experimental system, as constant activity can be measured from day 2 onwards. Overall, the SW system enables an effective method to assess simultaneously multiple resources through different indicators. Due to the possibility of representing the decisions and choices of the animals with different measurement methods, this system represents an extension of previous multi-choice tests. It is sufficiently sensitive to detect even small environmental changes and enables the collection of data on multifaceted aspects of resource use. To further investigate the importance of the resources for the animals adding an obstacle for reaching the resources will be advantageous (e.g. by adding distances between the resources). Moreover, the external validity of the approach can easily be extended by e.g. testing groups of animals and using animals of a variety of species, ages or sexes.

Aluminum monofluoride (AlF) has favorable properties for laser cooling and trapping experiments. The strong A¹Π ← X¹Σ⁺ band near 227 nm and the narrow a³Π ← X¹Σ⁺ band near 367 nm have rotationally and vibrationally closed transitions, making them well-suited for efficient photon scattering. Among the states relevant to laser cooling, the a³Π state is the least characterized, as its differing spin multiplicity from the ground state makes it difficult to access experimentally. This thesis reports on the spectroscopic characterization of the a³Π state and higher triplet states of AlF. Special emphasis is placed on the interpretation of anomalies in the molecular spectra, which occur due to interactions between different electronic states. The presented experiments are performed on supersonically cooled AlF molecules in a molecular beam. Pulsed and continuous-wave dye lasers (270–570 nm) and an RF transmission line (1–500 MHz) are used for resonant excitation.

The rotationally resolved spectra of the spin-forbidden a³Π ← X¹Σ⁺ band reveal that this band is weakly allowed due to the spin–orbit interaction of the a³Π state with distant singlet states. The vibrational, rotational, and spin–orbit coupling parameters of the a³Π, v = 0–7 states are extracted from these spectra. This allows the construction of an accurate electronic potential for the a³Π state and determination of the Franck–Condon matrix of the a³Π – X¹Σ⁺ system. All Λ-doublet transitions within the a³Π₁, v = 5, J = 1 level are recorded, and the hyperfine parameters of this level are determined. The observation of a spectral line with a full width at half maximum of 1.27 kHz, belonging to a hyperfine transition within the a³Π₁, v = 0 state, confirms the absence of lifetime broadening for transitions between two metastable levels.

The lowest rotational levels of the A¹Π, v = 6 and b³Σ⁺, v = 5 states are nearly isoenergetic and interact via spin–orbit coupling. These strongly mixed levels provide a doorway between the singlet and triplet manifolds. These perturbed levels are characterized from hyperfine-resolved spectra. The lifetimes of selected hyperfine levels, which range from 2 ns to 200 ns depending on the degree of singlet–triplet mixing, are determined by time-delayed ionization, Lamb-dip spectroscopy, and precise analysis of transition lineshapes.

Two theoretically predicted triplet states of AlF, the counterparts of the well-characterized D¹Δ and E¹Π states, had not been experimentally identified prior to this work. This thesis reports on the first characterization of the d³Π (v = 0–6) and e³Δ (v = 0–2) states, confirming the predicted energetic ordering of these states. In addition, the f³Σ⁺ (v = 0–2) states are characterized. Interestingly, the spectra of the d³Π, v = 3 ← a³Π, v = 3 band show an intensity distribution that is clearly different from that of all other diagonal bands. This band gets its weak, unexpected rotational structure due to intensity borrowing from the nearby e³Δ, v = 2 ← a³Π, v = 3 band. Analysis of this effect allows for the detailed characterization of the spin–orbit and spin–rotation interactions between the d³Π and e³Δ states.

Inflammation is a biological response to tissue injury or infection, mediated by signalling cascades and triggered by endogenous and exogenous danger signals. The nuclear protein High Mobility Group Box 1 (HMGB1), when released extracellularly, activates inflammatory pathways (e.g., nuclear factor kappa-light-chain-enhancer of activated B-cells (NF-κB) pathway) via pattern recognition receptors (PRRs) such as the Receptor for Advanced Glycation Endproducts (RAGE). Its activity is shaped by the positively charged A- and B-boxes and negatively charged acidic C-terminal tail, influencing receptor binding. Here, the anti-inflammatory properties of dendritic polyglycerol sulfate (dPGS), a synthetic, Heparin-mimetic polyanion, were investigated. Surface plasmon resonance (SPR) revealed that commercial HMGB1 from mammalian cells binds dPGS with nanomolar affinity (KD = 94.3 ± 1.7 nM). Alongside, recombinant full-length HMGB1 and subdomains expressed in E. coli and purified by a two-step affinity chromatography, showed that the A-box (KD = 83.5 nM) and A-+B-box (KD = 82.3 nM) exhibited similar affinities, while full-length HMGB1 displayed biphasic kinetics. Fitting with the heterogenous ligand binding model resulted in KD1 = 5 nM and KD2 = 0.009 nM, while dPGS also bound soluble (s)RAGE (KD1 = 6 nM, KD2 = 0.006 nM). HMGB1 bound sRAGE with KD1 = 70 nM, KD2 = 52 nM. Competitive binding studies demonstrated that dPGS impaired the HMGB1/sRAGE complex (IC50 = 3.3 nM). At higher dPGS concentrations biophysical data indicated a ternary complex of HMGB1/sRAGE/dPGS. Further, imaging showed that dPGS reduced lipopolysaccharide (LPS)-induced HMGB1/RAGE complex formation by 18 % in RAW 264.7 macrophages and 24 % in human microglia. NF-κB nuclear translocation induced by either HMGB1 or LPS was reduced to approximately 13 % of the control. As a result, RT-PCR showed that TNF-α mRNA levels were lowered owing to dPGS by 91 % (HMGB1 stimulus) and 81% (LPS stimulus), IL-6 by 71 % and 26 %, and MCP-1 by 34 % and 48 %, respectively. Protein quantification via ELISA confirmed reduced secretion levels of TNF-α upon co-stimulation with HMGB1 (50 %) or LPS (25 %). To prove that dPGS is capable to act intracellularly, uptake studies were performed with a cyanine 5 (Cy5) fluorescent label. Flow cytometry analyses showed time-dependent dPGS-Cy5 uptake peaking at 16–24 h. HMGB1 and LPS reduced intracellular accumulation by up to 30-40 % from 6 h onwards. Summarising, these results show that dPGS inhibits HMGB1-mediated inflammation at multiple structural and signalling checkpoints, supporting its potential as a therapeutic agent.

Maternal smoking during pregnancy is the most important avoidable cause of gestational morbidity and mortality, exerting permanent impacts on fetal health. This behavior is causally associated with fetal growth restriction (FGR), fetal malformations, and an increased vulnerability to disease later in life. The placenta is the functional interface between mother and fetus during pregnancy and is essential for healthy fetal development. This transient organ controls nutrient and gas exchange in-utero and secretes pregnancy modulating hormones. Molecular mechanisms that disturb the early development and function of the placenta by smoking remain largely unknown. This thesis uses a multi-omics approach to elucidate the phenotype-specifc molecular effects of maternal smoking on the developmental phase of the early human placenta. To this end, I evaluate tissue gene expression via single nuclei RNA sequencing (n = 5 smokers, n = 6 controls) and tissue relative protein abundances using ultra-sensitive phenotype-specific proteomics (n = 3 smokers, n = 3 controls). Results are validated in an independent human cohort (n = 23 smokers, n = 15 non-smokers), and functionally by in-vitro cigarette smoke component exposure using a primary human trophoblast stem cell model. Smoking status stratification for all cohorts was done via maternal plasma cotinine levels, a stable metabolite of nicotine and clinical gold standard. Results demonstrate that the syncytiotrophoblast (STB), the functional barrier of the placenta in direct contact with maternal blood, is profoundly affected and constitutes the only phenotype with compositional differences between smokers and non-smokers. The molecular profiling of the STB at a high resolution has been seldom performed, as its syncytial multinucleated nature poses challenges for its isolation. Trophoblast differentiation modelling demonstrates a preferential differentiation towards the STB lineage in smokers, in agreement with STB nuclei abundance increases in smokers. There are increases in pro-angiogenic secreted placental growth factor (PGF) gene expression (p < 0.001) with systemic relevance, as PlGF protein is increased in the blood of maternal smokers (p = 0.03). Smoking affected genes and proteins associated with cellular signalling, transmembrane transport, cellular stress and extracellular matrix remodelling. Molecular mechanisms associated with dysregulated markers at the transcriptomic and proteomic levels include increases in oxidative stress, activation of the xenobiotic detoxification aryl-hydrocarbon receptor pathway and disturbed mitochondrial energy metabolism capacity. Significantly higher numbers of mitochondrion are present in the STBs of smokers, as evaluated by immunofluorescence staining. Identified pathways can deregulate biological processes important for placental development and function, consequently impacting fetal growth and development. Beyond smoking, these findings provide a framework for understanding modulators of placental dysfunction that contribute to FGR – a leading cause of neonatal mortality and fetal morbidity – early in pregnancy, before the onset of clinical symptoms. I anticipate this data to be instrumental towards advancing FGR diagnostics and therapies to enhance pregnancy outcomes.

Dass Lernenden im Chemieunterricht häufig die sprachlichen Mittel fehlen, um komplexe Inhalte zu erfassen und wiederzugeben, stellt eine Herausforderung heutiger Schulbildung dar. In der vorliegenden Arbeit wird der Disaggregate-Instruction-Ansatz empirisch untersucht. Ziel des Ansatzes ist es, fachbezogene Kompetenzen von Schüler:innen dadurch zu fördern, dass Fachinhalte und Fachsprache zunächst getrennt voneinander vermittelt werden.

In einer vergleichenden Interventionsstudie wurde anhand einer exemplarisch konzipierten Unterrichtsreihe mit N = 464 Lernenden aus 20 Klassen der Sekundarstufe I untersucht, wie sich der Disaggregate-Instruction-Ansatz auf den Erwerb von fachlichen und kommunikativen Kompetenzen auswirkt. Besondere Aufmerksamkeit galt dabei Schüler:innen, deren bildungssprachliche Kompetenzen verhältnismäßig niedrig sind.

Die Ergebnisse zeigen, dass Lernende durch eine entkoppelte Vermittlung von Fachinhalten und korrespondierender Fachterminologie neue Konzepte vor allem dann besser erwerben, wenn sie über geringe bildungssprachliche Kompetenzen verfügen. Damit liefert die Dissertation Impulse für eine gerechtere und lernförderliche Gestaltung von Chemieunterricht in zunehmend heterogenen Klassenzimmern. So kann es in sprachlich schwächeren Lerngruppen sinnvoll sein, bei der Einführung naturwissenschaftlicher Konzepte neue Fachtermini zunächst auszuklammern.

This dissertation examines the mediatization and digitization of Islam, with a focus on Moroccan Muslims and mosque associations in Germany. Following a qualitative research design, it explores how digital technology and online media are used by Moroccan Muslim communities. It also examines how the construct of religious authority is perceived and embodied within this specific local context. By doing so, it aspires to identify the possible implications of digital media use on the dynamics of religious authority. Drawing on several case studies, hybrid ethnography, and semi-structured interviews with different actors within these Moroccan mosque associations, namely the associations’ administration, the imams (religious leaders), and the mosque community members, the research uncovers valuable insights regarding the digitization of Islam and Islamic authority in the local context. To do that, the study is guided by the mediatization and digitization of religion scholarship, which provides a framework for analyzing the interplay between the digitization processes and the dynamics of religious authority. The findings captured an increase in the use of digital media among the Moroccan Muslim associations under study, especially during the COVID-19 pandemic. Still, such development was unpredictable, inconsistent, and mostly dependent on offline circumstances. While the communities in focus tended to converge in resorting to the online sphere, they diverged in how they used digital technology and online media to achieve different purposes. The study also provided evidence for the contextual and dynamic nature of religious authority. By approaching the latter from a relational perspective, the interviews revealed that the imam can be considered the face of religious authority among Moroccan Muslim communities in the diasporic context. Such an authority draws mainly on the imam’s direct connection with the mosque community and his multi-dimensional role within the association’s setting. The interviews further suggested the importance and relevance of the construct of religious authority for Muslim communities, in terms of shaping religion-related decisions in this context. In addition to that, the study confirms that the digitization of religion has encouraged non-expert Muslim individuals to take their religiosity into their own hands and, therefore, actively engage in negotiating and constructing religious authority. Yet, it argues for the persistence of the traditional religious authority of the imam, given the nature of these local Muslim communities in the diasporic context. Thus, the study offers theoretical contributions by advancing our understanding of the mediatization and digitization of Islam in today’s digital age and by offering various insights into the concept of religious authority in this specific setting.

Die Einbindung von Frauen in den Arbeitsmarkt und die damit verbundene Erosion traditioneller Rollen- und Aufgabenverteilungen gehören zu den zentralen Merkmalen gegenwärtigen sozialen Wandels und verweisen ohne Zweifel auf neue Mechanismen der Verteilung von Lebenschancen. Die Tatsache, dass dem rasanten Anstieg weiblicher Erwerbsquoten eine Schlüsselrolle im Umbruch von Familie, Ökonomie und Wohlfahrtsstaat zukommt, ist dabei weniger umstritten als die sozialtheoretische Deutung der Entwicklung. Die vorliegende Arbeit folgt dem Ziel, die Auflösung ehemals separater Geschlechtersphären anhand der weiblichen Arbeitsmarktintegration in den USA historisch nachzuvollziehen und gesellschaftstheoretisch einzuordnen. Der Fall USA steht dabei für eine von Paradoxien und Ambivalenzen geprägte Entwicklung, die sich sowohl als Ausdruck eines individualistischen Liberalismus sowie als exemplarisch für eine allgemeine neoliberale Wende innerhalb moderner westlicher Wohlfahrtsstaaten verstehen lässt. Die übergeordnete Fragestellung lautet dabei, in welcher Weise sich die Strukturprinzipien gesellschaftlicher Ungleichheiten in den USA aufgrund der beschriebenen Verschiebungen im Geschlechterverhältnis verändert haben – und ob der Geschlechtszugehörigkeit überhaupt noch eine Bedeutung für die Verteilung von Lebenschancen zugeschrieben werden kann. Anhand einer empirischen Bestandsaufnahme (Teil I), der Gegenüberstellung und Evaluation zwei dominanter Deutungsmodelle (Teil II) sowie entlang der Revision gegenwärtiger Analyseinstrumentarien der feministischen Theorie (Teil III) kommt die Arbeit schließlich zu dem Schluss, dass sich gegenwärtige Geschlechterverhältnisse in den USA nur mit dem Blick auf ihre Verschränkung mit weiteren Ungleichheitslagen, vor allem der sozialen Klasse, analytisch nachvollziehen lassen.

The eastern section of the lower causeway and a small area of the south wall of the harbour area of the Bent Pyramid at Dahshur were archaeologically investigated between 2009 and 2018. A total of 424.6 kg of ceramic vessels was recovered, mostly in fragmentary condition and only rarely well-preserved or complete. The finds derived either from aeolian or fluvial sand layers, or from architectural contexts.

A fundamental aim of the present work was the creation of a fabric description system for the Dahshur pottery. At the outset of the study, the geomorphological conditions in the working area had to be examined in more detail. This knowledge formed the basis for the description system of the Dahshur pottery, modeled on the Vienna System. Another insight gained from the geology of the Dahshur region is the identification of a marl clay deposit at the mouth of the Sneferu Valley. In addition, it was established that both the Bent Pyramid and the Red Pyramid were founded on Pleistocene deposits (fluvial gravels and sands), which may have led to a different architectural layout of the two pyramids compared to the pyramids in Giza. In this regard, further important research remains to be carried out, as too little is yet known about the thickness of the Pleistocene sediments.

The core of the study consisted in the analysis of ceramic forms, particularly from a chronological perspective. Special attention was paid to vessel types whose dating has so far been difficult to fix in time, such as beer jars and miniature offering bowls. For the beer jars, a development from wide to narrow forms could be traced; for the miniature offering bowls, from large to small. For the latter ceramic type, the increased use of marl clay can be attested in the late Old Kingdom. In the study of beer jars, it was also observed that the Nile silt coatings required for this coarse ware resulted in a volume reduction of approximately 20%.

Another important topic of this work was dating the construction of the lower causeway and the harbour area, as well as their silting-up process. Both structures can be dated on the basis of the ceramic typology to the early to mid-4th Dynasty. Two later construction phases were identified at the lower causeway: a raising of the outer walls probably took place no later than in the reign of Niuserre (mid-5th Dynasty), and the insertion of a vault likely during the reign of Pepi I (6th Dynasty). The elevation of the outer walls of the lower causeway is associated with increasing landscape transformation, particularly heightened sand mobility. In the New Kingdom, with the decline of the Amarna period (ca. 1300 BCE), the cult at Sneferu’s valley temple ceased, and the dismantling of the temple — perhaps even of the entire pyramid complex — began.

The quantitative analysis of vessel types from excavation trench 1 showed that, in the aeolian sand layers of the 5th Dynasty, beer jars, miniature offering vessels, and, to a lesser extent, bread molds are predominant. This pottery was located primarily north of the lower causeway and, due to its association with hearths, faience beads, and intentionally broken vessels interpreted as ritual acts, can be understood as an element of a cult site. For various reasons, it can be assumed that this cult shows parallels to the sun temples at Abusir and is connected to the agricultural calendar of Ancient Egypt, which was primarily influenced by the Nile inundation occurring in summer. The deposition of cult ceramics during the 5th Dynasty at the lower causeway is therefore most plausibly interpreted in the context of festivals celebrated during the annual flooding of the Nile and during the harvest months.

The examination of landscape change at the end of the Holocene humid phase revealed that the present climatic state had not yet been reached in the first half of the 3rd millennium BCE. A shift towards a hyperarid climate comparable to today’s landscape very likely began at the transition from the 4th to the 5th Dynasty. The particular interest in a Sneferu cult throughout the 5th Dynasty can be well explained against the background of advancing aridification and Sneferu’s self-presentation — already in his lifetime — as a guarantor of sufficiently high Nile inundations and the resulting abundant harvests.

Diese Arbeit untersucht die Verbreitung von römischem Glas im kontinentalen barbaricum von der Zeitenwende bis zum Beginn der Völkerwanderungszeit unter Verwendung eines neuen Werkzeugs zur wissenschaftlichen Analyse umfangreicher Datenbanken, dem Digital Atlas of Innovations. Durch die Zusammenstellung einer gesamten Fundgattung auf (barbarisch-) europäischer Ebene wurde deutlich, dass es sich nicht um einen einheitlichen, linearen Entwicklungsprozess handelte, und es wurde nach Möglichkeiten gesucht, die unterschiedlichen Phänomene in einen sinnvollen Zusammenhang einzuordnen. Es wird argumentiert, dass der erste größere Niederschlag von Hohlgläsern im Fundgut des barbarischen Mittel- und Nordeuropas zwischen Rhein und Weser in die erste Hälfte des 1. Jahrhundert n. Chr., im Zuge der sogenannten „Okkupationszeit“, datiert, jedoch keinen besonderen Einfluss auf die Entwicklung der Glasproduktion im barbaricum genommen habe. Eine etwaige Entwicklung und Innovation im Bereich der Glasherstellung und -verarbeitung kann nur anhand von mehreren „Verständnisebenen“ stichpunktartig festgestellt werden. Drüber hinaus wurden zwei weitere Schwerpunktverschiebungen festgestellt, zum einen vom Rhein hin zur Donau im Zuge der „Markomannenkriege“ (166–180 n. Chr.) und zum anderen im 3. Jahrhundert n. Chr., sowohl von der Donau in den pontischen Raum, wo vor allem Becher mit eingeschliffenen Facetten über die großen osteuropäischen Flusssysteme bis in den skandinavischen Raum gelangten und so enge Kontakte über den Limes hinweg belegen, als auch eine Rückverschiebung in die ober- und niedergermanischen Provinzen sowie nach Raetien. Erst im aufkommenden frühen Mittelalter sind verschiedene Orte im ehemaligen barbaricum als regionale Zentren der Glasherstellung und -verarbeitung sicher ansprechbar.

The present study investigates the protein-polyelectrolyte complex formation and folding of proteins with a primary focus on how external factors such as temperature, salt concentrations, types of salt ions, reducing agents and pH affect the thermodynamic parameters. Using an array of biophysical techniques, including ITC, nanoDSF, DLS, DSC, CD and pressure-dependent FT-IR, this study provides a comprehensive thermodynamic perspective on protein stability and protein interactions with polyelectrolytes. To further quantify these protein-polyelectrolyte interactions and characterize protein behaviour under varying conditions, theoretical models based on the experimental data were validated and applied. The first part of this thesis contributes to the understanding of how ion-specific effects influence the thermodynamics of protein-polyelectrolyte interactions. The interaction of heparin and lysozyme in the presence of KGlu under varying salt concentrations was measured by ITC. The free energy of binding ΔGb depends strongly on salt concentration cs, with counterion release identified as the primary driving force for the interaction. In this investigation, a model for the free energy of binding ΔGb as the function of the two decisive variables temperature T and salt concentration cs was applied. This model highlights the significant role of hydration in the binding process. It was shown that temperature T, salt concentration cs and the type of ions present in the solution influence the free energy of binding ΔGb. In the second part, a critical comparison of Differential Scanning Fluorimetry nanoDSF and Differential Scanning Calorimetry DSC based on the thermal unfolding of lysozyme is presented. The unfolding of proteins is a well-understood problem, widely studied by thermal analysis such as Differential Scanning Calorimetry DSC, which needs high amounts of protein together with large protein concentrations. On the other hand, Differential Scanning Fluorimetry nanoDSF offers the advantage of requiring only small quantities and volumes of protein. For the first time, this study presents an evaluation model of the degree of unfolding α obtained from the nanoDSF fluorescence signal in order to determine the thermodynamic parameters. The thermal unfolding of the model protein lysozyme was measured at varying ranges of pH values and compared with the literature. Subsequently, the evaluation protocol of the nanoDSF fluorescence signal was applied to study the thermal stability of HMGB1, a redox-sensitive protein involved in numerous biological functions. NanoDSF enabled the determination of thermodynamic parameters with minimal protein sample requirements. Here, the impact of the disulfide bridge between Cys23 and Cys45 on the stability of the HMGB1 protein was explored. Several methods, like Differential Scanning Fluorimetry nanoDSF, fluorescence spectroscopy FT-IR, circular dichroism CD, and high-pressure FT-IR, were employed to quantify the unfolding and folding transitions. To understand its possible contribution to the stability and flexibility of HMGB1. The variants of the HMGB1 protein A-Box, B-Box, (A+B)-Box and a full-length HMGB1 were analysed. Moreover, the stability and flexibility of HMGB1 for the different redox states of A-Box and full-length HMGB1 were proved. Those different thermodynamic fingerprints of HMGB1 appear to have biological consequences, defining its function, as the redox state is closely connected to it and depends on the surrounding environment of the protein. Depending on its redox state, the HMGB1 protein changes the conformation and has distinct roles in the organism.

Soft tissue sarcomas represent a rare and highly heterogeneous group of tumors that pose complex challenges in both diagnosis and therapy. These difficulties range from limited awareness and delayed detection to potential misdiagnoses and the selection of appropriate, subtype-specific therapeutic modalities. Although established treatment concepts exist, many are still based on older and less differentiated approaches. Insights and resources derived from more common tumor entities will be of great value in advancing the goal of personalized oncology within sarcoma management.

Current (pre-)clinical and translational research demonstrates a clear trend toward subtype-specific and individualized therapeutic strategies. Innovations such as the integration of immunotherapies, TCR-T cell technologies, and the identification of predictive biomarkers are expected to enable increasingly refined treatment approaches tailored to the unique tumor profile and clinical context of each patient.

A paradigm shift is also emerging in radiation oncology: hypofractionation aims to shorten treatment duration and improve patient satisfaction. At the same time, multimodal treatment concepts and recognition of subtype-specific radiosensitivity are expected to allow for individualized dose adjustments, thus overcoming the current “one-size-fits-all” approach. In the future, patient-derived 3D tumor models may serve as “avatars” to test personalized therapeutic strategies and directly translate experimental findings into clinical care.

Furthermore, radiation oncology research will play a key role in implementing stereotactic body radiotherapy (SBRT)—already established as an effective local-ablative alternative to surgical resection in other tumor entities—for the management of soft tissue sarcoma metastases.

Collectively, these developments are paving the way toward increasingly differentiated and personalized care for patients with soft tissue sarcomas, with the overarching goal of sustainably improving therapeutic outcomes.

The field of nanophotonics studies the interaction of light and matter at microscopic scales. It plays an important role in many modern technologies, such as efficient solar cells, fast and secure quantum communication devices, and many other applications. The development of these technologies in a primarily experimental environment is challenging. As such, nanophotonics is a computationally heavily involved discipline. It often involves numerical simulations, optimization algorithms, as well as advanced modeling techniques, that all complement each other.

This cumulative dissertation explores the application of a particular kind of machine learning surrogate model, Gaussian processes, to certain optimization and modeling problems that appear in nanophotonics. Gaussian processes---in the context of nanophotonics---are trained on results of numerical simulations and can afterwards be used to approximate these simulations. They can help reduce the computational impact because approximate solutions obtained using Gaussian processes are much cheaper than results from numerical simulations. In addition to an approximation of the training inputs, the predictions of a Gaussian process are also associated with a measure of the confidence in the approximation. We develop methods and approaches that rely on these properties, implement them as (typically) Python code, investigate their performance and properties, and finally also apply them to solve two particular problems that appear in nanophotonics.

First, we employ Gaussian processes to accelerate parameter reconstruction tasks in optical nanometrology that involve computationally expensive model functions. In addition to aiding in the reconstruction of the model parameters that best explain some experimental measurement, trained Gaussian processes are then further used to determine the confidence bounds of the found model parameters. We present the methodology and perform extensive benchmarks, measuring its performance with respect to other methods that have been used to perform parameter reconstructions. We additionally apply it to give insights into the impact of properly or improperly chosen numerical accuracy parameters during a parameter reconstruction effort.

Second, we employ Gaussian processes for the quantification and optimization of the robustness of nanophotonic devices under an assumed manufacturing uncertainty. Gaussian processes are well suited for this task due to their intrinsic ability to model uncertainties. We present approaches that use Monte Carlo methods to assess the robustness of design candidates using Gaussian processes. We apply these approaches to assess the manufacturability of a novel cavity design, as well as to optimize certain optical resonators in a way that incorporates manufacturing uncertainties, first on a comparatively small domain using a multi-objective function, and second on a much larger domain.

The methods we present here are developed and applied in the context of nanophotonics. Their applicability, however, is not limited to this field.

Stress is a fundamental biological and psychological process that can have profound effects on brain structure and function. In the face of stress, a complex neuroendocrinological cascade is activated, enabling the individual to respond rapidly and adaptively to the stressor. Unfortunately, chronic activation of this adaptive stress response can lead to permanent brain changes that increase an individual's vulnerability to a variety of mental and somatic health problems. The present dissertation project aims to elucidate how different forms of stress influence structural and functional brain correlates. The dissertation consists of three studies, each addressing different aspects of stress-related brain changes. The first two studies focus on structural brain changes associated with childhood trauma (CT) and CT-related psychopathologies. Specifically, these studies sought to disentangle the effects of CT and adult CT-related psychopathologies – post-traumatic stress disorder (PTSD) and borderline personality disorder (BPD) on brain volume and cortical thickness. The underlying idea is to identify CT-related structural brain changes that may be initially adaptive but increase the risk of adult psychopathology, and psychopathology-specific structural brain changes that may explain disorder-specific symptom clusters. The third study investigates the neuroendocrinological basis of stress-related brain function by examining the effects of norepinephrine and cortisol on resting-state functional connectivity (RSFC) of the hippocampus and amygdala. Study one assessed hippocampal, amygdala and total brain volumes in three groups of women exposed to CT: (1) healthy women with CT, (2) women with CT and PTSD, and (3) women with CT and BPD. The study aimed to determine whether CT alone or in combination with PTSD or BPD was associated with volumetric reductions in limbic regions commonly implicated in stress responses. Contrary to expectations, there were no significant differences in hippocampal, amygdala or total brain volume between the groups. However, self-reported severity of CT was negatively associated with total brain volume, and self-reported depressive symptoms correlated with hippocampal volume reduction. The results suggest that reduced total brain volume may serve as a transdiagnostic vulnerability factor associated with CT, whereas reduced hippocampal volume may be associated with a CT-related depressive phenotype. The lack of significant group differences suggests that volumetric brain changes are not disorder specific but rather reflect underlying aetiological (CT) and symptomatic processes. Study two examined whole-brain cortical thickness in the same cohort to investigate whether cortical integrity was affected by CT and whether specific reductions in cortical thickness might indicate a latent vulnerability for the development of psychopathology. Results showed that both healthy women with CT and women with CT and BPD had reduced cortical thickness in the right lingual gyrus compared to healthy controls. In addition, healthy women with CT had reduced cortical thickness in the left lateral occipital lobe. Women with BPD showed further reductions in frontal and cingulate cortical thickness, consistent with the hypothesis that these regions are involved in emotion regulation deficits in BPD. Women with PTSD did not show significant differences in cortical thickness compared to healthy controls. The results of the second study suggest that reductions in cortical thickness in visual brain regions are associated with the experience of CT and may contribute to an increased risk of adult psychopathology. In addition, reductions in cortical thickness in cingulate and frontal brain areas appear to be specifically associated with individuals with CT-related BPD. Results from the first two studies suggest that CT is associated with reductions in total brain volume and cortical thickness in the lateral occipital lobe and lingual gyrus. While reduced cortical thickness in the lateral occipital lobe may function as an adaptive response that reduces trauma-related visual memory and reduces the risk of psychopathology, reductions in total brain volume and lingual gyrus thickness are likely to serve as vulnerability factors that increase susceptibility to mental disorders. In addition, disorder-specific structural changes were identified, with reductions in hippocampal volume associated with depressive symptoms and cortical thinning in frontal and cingulate regions appearing specific to BPD. Structural brain changes may represent both risk and resilience factors, with their role in psychopathology depending on the brain regions involved and their functional consequences. Study three focused on neuroendocrinological mechanisms linking stress and brain function in healthy male participants. This study used a pharmacological manipulation approach to investigate the independent and combined effects of norepinephrine and cortisol on hippocampal and amygdala RSFC in healthy young men. While no significant effects of either norepinephrine or cortisol alone on RSFC were observed, simultaneous elevation of both hormones led to increased connectivity between the hippocampus, amygdala, and cerebellum. This finding suggests that stress-related neuroendocrine responses may influence functional interactions between limbic structures and the cerebellum, possibly activating a functional network that is crucial for emotional memory processes. This dissertation offers valuable new insights into the relationship between stress and brain structure and function. Future research should build on these findings by enhancing generalizability, utilizing transdiagnostic and symptom-based approaches, empirically testing the relationship between brain changes and behavioral correlates, and further exploring the role of the cerebellum in stress and trauma.