Anders Malthe-Sørenssen

• Camposano, Anthony Val Canillas; Hafreager, Anders; Malthe-Sørenssen, Anders & Sveinsson, Henrik Andersen (2026). Silica-water model using the Vashishta force field. Journal of Chemical Physics. ISSN 0021-9606. 165(10). doi: 10.1063/5.0343021. Full text in Research Archive
• Schöb, Franziska Jessica; Binder, Alexander; Zamarian, Valentina; Sordi, Valeria; Scholz, Hanne & Malthe-Sørenssen, Anders [Show all 7 contributors for this article] (2026). Deep learning for predicting stem cell efficiency for use in beta cell differentiation. Scientific Reports. 16(1). doi: 10.1038/s41598-026-42830-3. Full text in Research Archive Show summary

  Recent clinical trial data show curative potential of cell therapy for diabetes, however the cells required are a bottleneck. Cell differentiation exhibits substantial variability, even among clones of stem cells generated from the same patient. Human experts struggle to see the difference between highly- and lowly-efficient cell clones early. We therefore propose an image-based deep learning model to guide the selection of the most efficient clones. We apply different deep learning models to learn the morphological differences between good and bad stem cell clones and classify them based on phase-contrast imaging. To gain insight into the learned features, we use layer-wise relevance propagation, and Fourier-based frequency analysis. Using an EfficientNet-V2-S model, we obtain a novel early prediction for the outcome of the differentiation process from patient-derived stem cells to β -cells using imaging. Clone level accuracy is 96.7 % at 53 hours after start of differentiation. The analysis of learned features shows that the structure of the cell population is an important predictive feature. This study is a proof-of-concept that deep learning combined with label-free imaging can be highly predictive and guide selection of stem cell clones, thereby reducing cost of β -cell production.

• Nymoen, Ingeborg; Rogge, Frederik Sebastian; Aasen, Anna Hjertvik; Grødem, Sverre; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle [Show all 9 contributors for this article] (2026). Task and Behavior-Related Variables Are Encoded by the Postrhinal and Medial Entorhinal Cortex During Non-Spatial Associative Learning. eLIFE. 14. doi: 10.7554/elife.106095.1. Full text in Research Archive Show summary

  The medial entorhinal cortex (MEC) is pivotal in spatial computations and episodic memory. In particular, an animal’s position can be decoded from the activity of entorhinal grid cells. However, it remains elusive whether MEC could play a more general role in different types of associative learning and how the representations develop during the learning process. It has been shown that the postrhinal cortex (POR), which is directly connected to MEC, integrates visual stimuli with salient outcomes. Here, we use a non-spatial visual association task to investigate whether MEC neurons represent low-level visual cues during learning. Using a Go/NoGo visual association task, we recorded neural activity in MEC and POR throughout the learning phase as mice associated drifting gratings with rewarded, aversive, or neutral outcomes. Our findings reveal that the neural tuning curves in both the POR and MEC change with the learning of the task. From the start of training, the POR neurons exhibited response tuning to the visual cues, and the tuning was stable to cue orientations during learning. In contrast, MEC neurons did not initially respond very strongly to visual cues but developed a robust tuning toward the rewarded trials. While the MEC’s representation of visual information was limited, it encoded other task elements. A large fraction of the neurons formed distinct functional clusters that were either activated or suppressed by reward-related behavior. Remarkably, these clusters segregated anatomically in MEC and maintained strong within-cluster correlations before and after training. Notably, although the same functional clusters were apparent in the POR, they did not show any anatomical structure as in the MEC. Task reversal induced significant changes in network responses across both regions, with a decrease in overall task-responsive neurons but a slight increase in stimulus representation. Strikingly, information about the choice to lick emerged with learning in both brain areas, and most significantly within the functional cell clusters representing reward consumption and plus-cue stimulus. Our results demonstrate that although neurons in MEC and POR develop behavior-modulated tuning during learning of a non-spatial visual association task, the MEC exhibits stronger within-cluster correlations and anatomical organization. Conversely, the POR population exhibits less structural organization and more specific stimulus-tuning, which is reflective of being a higher visual association area. Our findings reveal that the MEC can encode task– and behavior-related variables beyond spatial information.

• Pettersen, Markus Borud; Schøyen, Vemund Sigmundson; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2025). Learning place cells and remapping by decoding the cognitive map. eLIFE. 13. doi: 10.7554/elife.99302. Full text in Research Archive Show summary

  Hippocampal place cells are known for their spatially selective firing and are believed to encode an animal’s location while forming part of a cognitive map of space. These cells exhibit marked tuning curves and rate changes when an animal’s environment is sufficiently manipulated, in a process known as remapping. Place cells are accompanied by many other spatially tuned cells, such as border cells and grid cells, but how these cells interact during navigation and remapping is unknown. In this work, we build a normative place cell model wherein a neural network is tasked with accurate position reconstruction and path integration. Motivated by the notion of a cognitive map, the network’s position is estimated directly from its learned representations. To obtain a position estimate, we propose a non-trainable decoding scheme applied to network output units, inspired by the localized firing patterns of place cells. We find that output units learn place-like spatial representations, while upstream recurrent units become boundary-tuned. When the network is trained to perform the same task in multiple simulated environments, its place-like units learn to remap like biological place cells, displaying global, geometric, and rate remapping. These remapping abilities appear to be supported by rate changes in upstream units. While the model does not learn grid-like units, its place unit centers form clusters organized in a hexagonal lattice in open fields. When we decode the center locations of CA1 place fields in mice, we find preliminary evidence of a similar clustering tendency. This suggests a potential mechanism for the interaction between place cells, border cells, and grid cells. Our model provides a normative framework for learning spatial representations previously reserved for biological place cells, providing new insight into place cell field formation and remapping.

• Beshkov, Kosio & Malthe-Sørenssen, Anders (2025). [2509.24895] Towards Understanding the Shape of Representations in Protein Language Models. arXiv. Full text in Research Archive Show summary

  While protein language models (PLMs) are one of the most promising avenues of research for future de novo protein design, the way in which they transform sequences to hidden representations, as well as the information encoded in such representations is yet to be fully understood. Several works have attempted to propose interpretability tools for PLMs, but they have focused on understanding how individual sequences are transformed by such models. Therefore, the way in which PLMs transform the whole space of sequences along with their relations is still unknown. In this work we attempt to understand this transformed space of sequences by identifying protein structure and representation with square-root velocity (SRV) representations and graph filtrations. Both approaches naturally lead to a metric space in which pairs of proteins or protein representations can be compared with each other. We analyze different types of proteins from the SCOP dataset and show that the Karcher mean and effective dimension of the SRV shape space follow a non-linear pattern as a function of the layers in ESM2 models of different sizes. Furthermore, we use graph filtrations as a tool to study the context lengths at which models encode the structural features of proteins. We find that PLMs preferentially encode immediate as well as local relations between residues, but start to degrade for larger context lengths. The most structurally faithful encoding tends to occur close to, but before the last layer of the models, indicating that training a folding model ontop of these layers might lead to improved folding performance.

• Camposano, Anthony Val Canillas; Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Genetic Algorithm Workflow for Parameterization of a Water Model Using the Vashishta Force Field. Journal of Physical Chemistry B. ISSN 1520-6106. 129(4), p. 1331–1342. doi: 10.1021/acs.jpcb.4c06389. Full text in Research Archive Show summary

  Water participates in countless processes on Earth, and the properties of mineral surfaces can be drastically changed in the presence of water. For example, the fracture toughness of silica glass is reduced by 25% for water-filled cracks than for dry cracks [Geophys. Res.: Solid Earth 2018, 123, 9341–9354]. An accurate description of water is therefore essential for modeling the behavior of minerals in aqueous environments and, in particular, for modeling dynamic processes such as fracture, where the mechanical response of water may play an important role. On the molecular scale, molecular dynamics simulations with empirical force field methods provide a way to study large molecular systems at a relatively low computational cost. Many water models have been developed previously; however, a computationally cheap water model capable of describing reactions with minerals is lacking. Here, we present a parametrization of the water potential using the Vashishta potential form [Phys. Rev. B 1990, 41, 12197–12209]. For this 3-point water model, we obtain good agreement with experimental transport and liquid–vapor properties. Importantly, the Vashishta form opens up compatibility with existing silica glass models, thus enabling the simulation of mineral–water interactions.

• Holzhausen, Konstantin; Haga, Mia Merlid; Torvik, Håkon Olav; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2025). Exploring biologically inspired mechanisms of adversarial robustness. Neural Computing & Applications. ISSN 0941-0643. 37, p. 9331–9342. doi: 10.1007/s00521-025-11019-6. Full text in Research Archive Show summary

  Backpropagation-optimized artificial neural networks, while precise, lack robustness, leading to unforeseen behaviors that affect their safety. Biological neural systems do solve some of these issues already. Unlike artificial models, biological neurons adjust connectivity based on neighboring cell activity. Understanding the biological mechanisms of robustness can pave the way toward building trustworthy and safe systems. Robustness in neural representations is hypothesized to correlate with the smoothness of the encoding manifold. Recent work suggests power law covariance spectra, which were observed studying the primary visual cortex of mice, to be indicative of a balanced trade-off between accuracy and robustness in representations. Here, we show that unsupervised local learning models with winner takes all dynamics learn such power law representations, providing upcoming studies a mechanistic model with that characteristic. Our research aims to understand the interplay between geometry, spectral properties, robustness, and expressivity in neural representations. Hence, we study the link between representation smoothness and spectrum by using weight, Jacobian and spectral regularization while assessing performance and adversarial robustness. Our work serves as a foundation for future research into the mechanisms underlying power law spectra and optimally smooth encodings in both biological and artificial systems. The insights gained may elucidate the mechanisms that realize robust neural networks in mammalian brains and inform the development of more stable and reliable artificial systems.

• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Velocity-Dependent Dynamics of Friction and Wear. Tribology letter. ISSN 1023-8883. doi: 10.1007/s11249-025-02045-5. Full text in Research Archive
• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2025). Tailoring Frictional Properties of Surfaces Using Diffusion Models. Journal of Physical Chemistry C. ISSN 1932-7447. 129(32), p. 14559–14564. doi: 10.1021/acs.jpcc.5c02768. Full text in Research Archive Show summary

  This work introduces an approach for precisely designing surface friction properties using a conditional generative machine learning model, specifically a denoising diffusion probabilistic model (DDPM). We created a dataset of synthetic surfaces with frictional properties determined by molecular dynamics simulations, which trained the DDPM to predict surface structures from desired frictional outcomes. Unlike traditional trial-and-error and numerical optimization methods, our approach directly yields surface designs meeting specified frictional criteria with high accuracy and efficiency. This advancement in material surface engineering demonstrates the potential of machine learning to reduce the iterative nature of surface design processes. Our findings provide a pathway for tailoring surface properties and suggest broader applications in materials science where surface characteristics are critical.

• Camposano, Anthony Val Canillas; Nordhagen, Even Marius; Malthe-Sørenssen, Anders & Sveinsson, Henrik Andersen (2025). An Extended Energy-Biased Aggregation-Volume-Bias Monte Carlo (EB-AVBMC) Method for Nucleation Simulation of a Reactive Water Potential. Journal of Chemical Theory and Computation. ISSN 1549-9618. 21(14), p. 6769–6776. doi: 10.1021/acs.jctc.5c00722. Full text in Research Archive
• Schøyen, Vemund Sigmundson; Beshkov, Kosio; Pettersen, Markus Borud; Hermansen, Erik; Holzhausen, Konstantin & Malthe-Sørenssen, Anders [Show all 8 contributors for this article] (2025). Hexagons all the way down: grid cells as a conformal isometric map of space. PLoS Computational Biology. ISSN 1553-734X. 21(2). doi: 10.1371/journal.pcbi.1012804. Full text in Research Archive Show summary

  Grid cells in the entorhinal cortex are known for their hexagonal spatial activity patterns and are thought to provide a neural metric for space, and support path integration. In this study, we further investigate grid cells as a metric of space by optimising them for a conformal isometric (CI) map of space using a model based on a superposition of plane waves. By optimising the phases within a single grid cell module, we find that the module can form a CI of two-dimensional flat space with phases arranging into a regular hexagonal pattern, supporting an accurate spatial metric. Additionally, we find that experimentally recorded grid cells exhibit CI properties, with one example module showing a phase arrangement similar to the hexagonal pattern observed in our model. These findings provide computational and preliminary experimental support for grid cells as a CI-based spatial representation. We also examine other properties that emerge in CI-optimised modules, including consistent energy expenditure across space and the minimal cell count required to support unique representation of space and maximally topologically persistent toroidal population activity. Altogether, our results suggest that grid cells are well-suited to form a CI map, with several beneficial properties arising from this organisation.

• Lensjø, Kristian; Nymoen, Ingeborg; Rogge, Frederik Sebastian; Tøndel, Kristin; Sugden, Arthur & Shurnayte, Inga [Show all 11 contributors for this article] (2025). Local inhibitory circuits mediate cortical reactivations and memory consolidation. Science Advances. 11(22), p. 1–14. doi: 10.1126/sciadv.adu9800. Full text in Research Archive Show summary

  Highly salient events activate neurons across various brain regions. During subsequent rest or sleep, the activity patterns of these neurons often correlate with those observed during the preceding experience. Growing evidence suggests that these reactivations play a crucial role in memory consolidation, the process by which experiences are solidified in cortical networks for long-term storage. Here, we use longitudinal two-photon Ca2+ imaging alongside paired LFP recordings in the hippocampus and cortex, to show that targeted manipulation of PV+ inhibitory neurons in the lateral visual cortex after daily training selectively attenuates cue-specific reactivations and learning, with only minute effects on spontaneous activity and no apparent effect on normal function such as visual cue–elicited responses during training. In control mice, reactivations were biased toward salient cues, persisted for hours after training had ended, and the prevalence of reactivations was aligned with the learning process. Overall, our results underscore a crucial role for cortical reactivations in memory consolidation.

• Jensen, Mikkel Metzsch; Malthe-Sørenssen, Anders & Sveinsson, Henrik Andersen (2025). Strong strain dependence of friction in graphene kirigami allows engineering a negative coefficient of friction. Proceedings of the National Academy of Sciences of the United States of America (PNAS). ISSN 0027-8424. 122(28). doi: 10.1073/pnas.2501728122. Full text in Research Archive Show summary

  The friction force typically increases linearly with normal load with a constant of proportionality called the coefficient of friction. Most materials exhibit a positive friction coefficient, so that an increase in the normal load leads to an increase in the friction force. Recently, materials with negative friction coefficients have been observed at meticulously constructed interfaces due to an interplay between superstructures at heterojunctions, out-of-plane buckling, or the ordering of thin water films. However, the magnitude of the negative friction coefficient is typically much less than 10^-2, the geometries are highly restrictive, and the mechanisms are difficult to scale to larger systems. Here, we show that a friction coefficient of -0.08 can be obtained for a graphene sheet modified with kirigami-inspired cuts when inducing sheet tension through a normal load coupling. We use molecular dynamics simulations to show that the frictional behavior of kirigami cut sheets exhibits a strong nonmonotonic dependence on in-plane strain, while only weakly influenced by normal loading. We argue that the strong influence of strain arises from changes in commensurability between the graphene sheet and the substrate when the sheet deforms and buckles out of plane. We propose a simple nanomachine design that couples normal loading and sheet tension which enables the realization of different frictional laws, including a negative coefficient of friction. This represents a unique approach to creating tunable frictional surfaces and opens up applications in sheet-like systems across scales.

• Combriat, Thomas Michel Daniel; Olsen, Petter Angell; Låstad, Silja Borring; Malthe-Sørenssen, Anders; Krauss, Stefan Johannes Karl & Dysthe, Dag Kristian (2024). Acoustic Wave-Induced Stroboscopic Optical Mechanotyping of Adherent Cells. Advanced Science. 11(16). doi: 10.1002/advs.202307929. Full text in Research Archive Show summary

  In this study, a novel, high content technique using a cylindrical acoustic transducer, stroboscopic fast imaging, and homodyne detection to recover the mechanical properties (dynamic shear modulus) of living adherent cells at low ultrasonic frequencies is presented. By analyzing the micro-oscillations of cells, whole populations are simultaneously mechanotyped with sub-cellular resolution. The technique can be combined with standard fluorescence imaging allowing to further cross-correlate biological and mechanical information. The potential of the technique is demonstrated by mechanotyping co-cultures of different cell types with significantly different mechanical properties.

• Odden, Tor Ole Bigton; Tyseng, Halvor; Mjaaland, Jonas Timmann; Kreutzer, Markus Fleten & Malthe-Sørenssen, Anders (2024). Using text embeddings for deductive qualitative research at scale in physics education. Physical Review Physics Education Research. 20(2). doi: 10.1103/PhysRevPhysEducRes.20.020151. Full text in Research Archive Show summary

  We propose a technique for performing deductive qualitative data analysis at scale on text-based data. Using a natural language processing technique known as text embeddings, we create vector-based representations of texts in a high-dimensional meaning space within which it is possible to quantify differences in meaning as vector distances. To apply the technique, we build off prior work that used topic modeling via latent Dirichlet allocation to thematically analyze 18 years of the Physics Education Research Conference Proceedings literature. We first extend this analysis through 2023. Next, we create embeddings of all texts and, using representative articles from the 10 topics found by the LDA analysis, define centroids in the meaning space. We calculate the distances between every article and centroid and use the inverted, scaled distances between these centroids and articles to create an alternate topic model. We benchmark this model against the LDA model results and show that this embeddings model recovers most of the trends from that analysis. Finally, to illustrate the versatility of the method, we define eight new topic centroids derived from a review of the physics education research literature by Docktor and Mestre and reanalyze the literature using these researcher-defined topics. Based on these analyses, we critically discuss the features, uses, and limitations of this method and argue that it holds promise for flexible deductive qualitative analysis of a wide variety of text-based data that avoids many of the drawbacks inherent to prior NLP methods.

• Ye, Bo; Malthe-Sørenssen, Anders & Jettestuen, Espen (2024). An adaptive mesh scheme of the lattice spring model based on geometrical continuity. Modelling and Simulation in Materials Science and Engineering. ISSN 0965-0393. 32(5). doi: 10.1088/1361-651X/ad4e50. Full text in Research Archive Show summary

  An adaptive mesh scheme is introduced for the lattice spring model (LSM), where the original triangular cells are subdivided into a set of smaller triangular cells. The scheme is based on geometrical continuity at the heterogeneous mesh boundary, where the refined grid cells intersect the original cell edge. The LSM simulations on the refined grid show a superior computational efficiency to the uniform grid. Each subdivision reduces the original cell edges by a factor of two. The refinement procedure was recursively applied ten times before any marked loss in accuracy was observed. The accuracy of the adaptive model is on par with a regular grid approach. More specifically, the characteristics of fracture cavity are comparable with a uniform grid of the same mesh density as the smallest cells in the adaptive approach. The fracture criterion such as J-integral, the elastic energy of the grid and potential energy change due to fracture growth and strain loading agree well with the theory of a mode I fracture, which enables simulations of process such as sub-critical fracture with a wide dynamic range.

• Schøyen, Vemund; Pettersen, Markus; Holzhausen, Konstantin; Fyhn, Marianne; Malthe-Sørenssen, Anders & Lepperød, Mikkel (2023). Coherently remapping toroidal cells but not Grid cells are responsible for path integration in virtual agents. iScience. Full text in Research Archive
• Ye, Bo; Jettestuen, Espen & Malthe-Sørenssen, Anders (2023). A node-splitting lattice spring model coupled with a J-integral formulation as a fracture criterion. International Journal of Fracture. ISSN 0376-9429. p. 1–19. doi: 10.1007/s10704-023-00749-0. Full text in Research Archive Show summary

  A global energy minimization criterion based on Griffith’s theory is introduced for the node-splitting lattice spring model. The fracture criterion is computed by both direct numerical simulations of energy release rate G and through a J-integral formulation for comparison and validation. For mode I fractures, the standard implementation of J-integral formulation yields very good estimations of the energy release rate, but for mixed mode fracture the estimations deviates from the direct calculated energy release rate. The reasons for this discrepancy are elucidated and an approach to best approximate the J value is given. This method is compared with the more standard maximum tip stress threshold crack criterion, and shows a much better prediction of the energy release rate and is more robust under grid refinement.

• Schøyen, Vemund Sigmundson; Pettersen, Markus Borud; Holzhausen, Konstantin; Fyhn, Marianne; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2023). Coherently remapping toroidal cells but not Grid cells are responsible for path integration in virtual agents. iScience. 26(11), p. 1–17. doi: 10.1016/j.isci.2023.108102. Full text in Research Archive Show summary

  It is widely believed that grid cells provide cues for path integration, with place cells encoding an animal’s location and environmental identity. When entering a new environment, these cells remap concurrently, sparking debates about their causal relationship. Using a continuous attractor recurrent neural network, we study spatial cell dynamics in multiple environments. We investigate grid cell remapping as a function of global remapping in place-like units through random resampling of place cell centers. Dimensionality reduction techniques reveal that a subset of cells manifest a persistent torus across environments. Unexpectedly, these toroidal cells resemble band-like cells rather than high grid score units. Subsequent pruning studies reveal that toroidal cells are crucial for path integration while grid cells are not. As we extend the model to operate across many environments, we delineate its generalization boundaries, revealing challenges with modeling many environments in current models.

• Nordhagen, Even Marius; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2023). Diffusion-Driven Frictional Aging in Silicon Carbide. Tribology letter. ISSN 1023-8883. 71(3), p. 1–11. doi: 10.1007/s11249-023-01762-z. Full text in Research Archive Show summary

  Abstract Friction is the force resisting relative motion of objects. The force depends on material properties, loading conditions and external factors such as temperature and humidity, but also contact aging has been identified as a primary factor. Several aging mechanisms have been proposed, including increased “contact quantity” due to plastic or elastic creep and enhanced “contact quality” due to formation of strong interfacial bonds. However, comparatively less attention has been given to other mechanisms that enhance the “contact quantity”. In this study, we explore the influence of crystal faceting on the augmentation of “contact quantity” in cubic silicon carbide, driven by the minimization of surface free energy. Our observations reveal that the temporal evolution of the frictional aging effect follows a logarithmic pattern, akin to several other aging mechanisms. However, this particular mechanism is driven by internal capillary forces instead of the normal force typically associated with friction. Due to this fundamental distinction, existing frictional aging models fail to comprehensively explain the observed behavior. In light of these findings, we derive a model for the evolution of contact area caused by diffusion-driven frictional aging, drawing upon principles from statistical mechanics. Upon application of a normal force, the friction force is increased due to plastic creep. This investigation presents an alternative explanation for the logarithmic aging behavior observed and offers the potential to contribute to the development of more accurate friction models. Graphical Abstract

• Najafi, Fahimeh; Sveinsson, Henrik Andersen; Dreierstad, Christer; Glad, Hans Erlend Bakken & Malthe-Sørenssen, Anders (2023). Modeling the relationship between mechanical yield stress and material geometry using convolutional neural networks. Applied Physics Letters. ISSN 0003-6951. 123(11). doi: 10.1063/5.0160338. Full text in Research Archive Show summary

  Machine learning methods can be used to predict the properties of materials from their structure. This can be particularly useful in cases where other standard methods for finding material properties are time and resources consuming to use on large sample spaces. In this work, we study the strength of α-quartz crystals with a porous layer created by simplex noise as the shape of porosity. We train a neural network to predict the yield stress of these systems under both shear and tensile deformation. Molecular dynamics simulations are used for a randomly selected sample of possible structures in order to generate the ground truth to be used as the training data. We employ deep convolutional neural networks (CNNs) which are commonly used when dealing with image or image-like data since the input data for the problem in hand are a binary 2D structure of the porous layer of the systems. The trained CNN can predict the yield stress of a system based on the geometry of that given system, with much higher precision compared to a baseline polynomial regression method. Saliency maps created with the trained model show that the model predictions are most sensitive to altering structures near high-stress regions, indicating that the model makes predictions based on reasonable physics.

• Guren, Marthe Grønlie; Sveinsson, Henrik Andersen; Malthe-Sørenssen, Anders & Renard, Francois (2022). Nanoscale Damage Production by Dynamic Tensile Rupture in α-Quartz. Geophysical Research Letters. ISSN 0094-8276. 49(20). doi: 10.1029/2022GL100468. Full text in Research Archive Show summary

  The creation of new fractures during an earthquake produces rock damage and contributes to the dissipation of strain energy. During dynamic rupture propagation, tensile microfractures can form in the earthquake process zone and in the domains around a fault that host large transient tensile stress. These microfractures can produce rock fragments with a wide range of sizes. Using molecular dynamics simulations, we model tensile rupture propagation in α-quartz under conditions of stress that occur during earthquake propagation. Our results show that for rupture speeds below 15% of the Rayleigh wave speed, fractures propagate linearly. At higher speeds, fracture propagation undergoes path instabilities with crack oscillations and microbranching leading to the formation of nanoscale roughness and fragments. This nanoscale damage can form in and around the earthquake process zone before any significant slip has occurred on the fault. The produced nanoparticles may control further energy dissipation during frictional slip.

• Odden, Tor Ole Bigton; Silvia, Devin W. & Malthe-Sørenssen, Anders (2022). Using computational essays to foster disciplinary epistemic agency in undergraduate science. Journal of Research in Science Teaching. ISSN 0022-4308. 60(5), p. 937–977. doi: 10.1002/tea.21821. Full text in Research Archive Show summary

  This article reports on a study investigating how computational essays can be used to help students in higher education STEM take up disciplinary epistemic agency—cognitive control and responsibility over one's own learning within the scientific disciplines. Computational essays are a genre of scientific writing that combine live, executable computer code with narrative text to present a computational model or analysis. The study took place across two contrasting university contexts: an interdisciplinary data science and modeling course at a large research university in the Midwestern United States, and a third-semester physics course at a large research university in Scandinavia. Over the course of a semester, computational essays were simultaneously and independently used in both courses, and comparable datasets of student artifacts and retrospective interviews were collected from both student populations. These data were analyzed using a framework that operationalized the construct of disciplinary epistemic agency across the dimensions of programming, inquiry, data analysis and modeling, and communication. Based on this analysis, we argue that computational essays can be a useful tool for fostering disciplinary epistemic agency within higher education science due to their combination of adaptability and disciplinary authenticity. However, we also argue that educational contexts, scaffolding, expectations, and student backgrounds can constrain and influence the ways in which students choose to take up epistemic agency.

• Hanssen, Kine Ødegård & Malthe-Sørenssen, Anders (2022). Perineuronal nets restrict transport near the neuron surface: A coarse-grained molecular dynamics study. Frontiers in Computational Neuroscience. 16. doi: 10.3389/fncom.2022.967735. Full text in Research Archive Show summary

  Perineuronal nets (PNNs) are mesh-like extracellular matrix structures that wrap around certain neurons in the central nervous system. They are hypothesized to stabilize memories in the brain and act as a barrier between cell and extracellular space. As a means to study the impact of PNNs on diffusion, the nets were approximated by negatively charged polymer brushes and simulated by coarse-grained molecular dynamics. Diffusion constants of single neutral and single charged particles were obtained in directions parallel and perpendicular to the brush substrate. The results for the neutral particle were compared to different theories of diffusion in a heuristic manner. Diffusion was found to be considerably reduced for brush spacings smaller than 10 nm, with a pronounced anisotropy for dense brushes. The exact dynamics of the chains was found to have a negligible impact on particle diffusion. The resistance of the brush proved small compared to typical values of the membrane resistance of a neuron, indicating that PNNs likely contribute little to the total resistance of an enwrapped neuron.

• Gløersen, Øyvind Nøstdahl; Colosio, Alessandro L.; Boone, Jan; Dysthe, Dag Kristian; Malthe-Sørenssen, Anders & Capelli, Carlo [Show all 7 contributors for this article] (2022). Modeling V̇o2 on-kinetics based on intensity-dependent delayed adjustment and loss of efficiency (DALE). Journal of applied physiology. ISSN 8750-7587. 132(6), p. 1480–1488. doi: 10.1152/japplphysiol.00570.2021. Full text in Research Archive Show summary

  This study presents and evaluates a new mathematical model of V̇o2 on-kinetics, with the following properties: 1) a progressively slower primary phase following the size-principle of motor unit recruitment, explaining the delayed V̇o2 steady state seen in the heavy exercise intensity domain, and 2) a severe-domain slow component modeled as a time-dependent decrease in efficiency. Breath-by-breath V̇o2 measurements from eight subjects performing step cycling transitions, in the moderate, heavy, and severe exercise domains, were fitted to the conventional three-phase model and the new model. Model performance was evaluated with a residual analysis and by comparing Bayesian (BIC) and corrected Akaike (AICc) information criteria. The residual analysis showed no systematic deviations, except perhaps for the initial part of the primary phase. BIC favored the new model, being 9.3 (SD 7.1) lower than the conventional model whereas AICc was similar between models. Compared with the conventional three-phase model, the proposed model distinguishes between the kinetic adaptations in the heavy and severe domains by predicting a delayed steady-state V̇o2 in the heavy and no steady-state V̇o2 in the severe domain. This allows to determine when stable oxygen costs of exercise are attainable and it also represents a first step in defining time-dependent oxygen costs when stable energy conversion efficiency is not attainable. NEW & NOTEWORTHY: We propose and assess a new minimalistic integrated model for the V̇o2 on-kinetics, inspired by the currently available best evidence of the underlying mechanisms. We show that the model provides a similar fit as the conventionally used three-phase model, even though a stricter data fitting method is used for the proposed model. The proposed model clarifies misconceptions related to the V̇o2 slow component’s behavior, by clearly predicting that steady-state V̇o2 is attainable in the moderate and heavy exercise intensity domains. Furthermore, the model opens new possibilities for assessing oxygen cost during severe intensity exercise without the fallible assumption of time-constant energy-conversion efficiency.

• Malthe-Sørenssen, Anders (2021). The onset of a slip. Nature Physics. ISSN 1745-2473. 17(9), p. 983–985. doi: 10.1038/s41567-021-01312-1. Full text in Research Archive
• Guren, Marthe Grønlie; Sveinsson, Henrik Andersen; Hafreager, Anders; Jamtveit, Bjørn; Malthe-Sørenssen, Anders & Renard, Francois (2021). Molecular dynamics study of confined water in the periclase-brucite system under conditions of reaction-induced fracturing. Geochimica et Cosmochimica Acta. ISSN 0016-7037. 294, p. 13–27. doi: 10.1016/j.gca.2020.11.016. Full text in Research Archive Show summary

  The volume-increase associated with hydration reactions in rocks may lead to reaction-induced fracturing, but requires a stable water film to be present at reactive grain boundaries even when subject to compressive stress. Hydration of periclase to brucite is associated with a solid volume increase of ca. 110%. Recent experiments on the periclase-brucite system observed that when the effective mean stress exceeds 30 MPa, the reaction rate slows down dramatically. We hypothesize that for the brucite forming reaction to progress, the fluid film between grains must remain stable. If the applied pressure becomes larger than the hydration force, the fluid film will collapse and be squeezed out of the grain contacts. To quantify this effect, we study the behavior of a water film confined between periclase or brucite surfaces subject to compressive stress, by performing molecular dynamics simulations. The simulations are carried out using the ClayFF force field and the single point charge (SPC) water model in the molecular dynamics simulations program LAMMPS. The setup consists of two interfaces of either periclase or brucite surrounded by water. Our simulations show that when the pressure reaches a few tens of MPa, the water film collapses and reduces the water film to one or two water layers, while the self-diffusion coefficient of water molecules by a factor of eight. A water film thickness below two water layers is thinner than the size of the hydration shell around Mg2+-ions, which will limit ion-transport. The observed collapse of the water film to a single layer at a normal pressure of 25–30 MPa might explain the observed slow-down of reaction-induced fracturing in the periclase-brucite system.

• Sveinsson, Henrik Andersen; Ning, Fulong; Cao, Pinqiang; Fang, Bin & Malthe-Sørenssen, Anders (2021). Grain-Size-Governed Shear Failure Mechanism of Polycrystalline Methane Hydrates. Journal of Physical Chemistry C. ISSN 1932-7447. 125(18), p. 10034–10042. doi: 10.1021/acs.jpcc.1c00901. Full text in Research Archive Show summary

  The shear failure mechanism of polycrystalline gas hydrates is critical for understanding marine geohazards related to gas hydrates under a changing climate and for safe gas recovery from gas hydrate reservoirs. Since current experimental techniques cannot resolve the mechanism on a spatial and temporal nanoscale, molecular simulations can assist with proposing and substantiating nanoscale failure mechanisms. Here, we report the shear failure of polycrystalline methane hydrates using direct molecular dynamics simulations. Based on these simulations, we suggest two modes of shear behavior, depending on the grain sizes, d, in the polycrystal: grain-size-strengthening behavior with a d1/3 grain size dependence for small grain sizes and grain-size-weakening behavior for large grain sizes. Through the crossover from strengthening to weakening behavior, the failure mode changes from shear failure with a failure plane parallel to the applied shear to tensile failure with a failure plane lying at an angle with the applied shear, spanning a network of grain boundaries. The existence of such a change in mechanism suggests that the Hall–Petch breakdown in methane hydrates is due to a change from grain boundary sliding to tensile opening being the most important failure mechanism when the grain size increases.

• Odden, Tor Ole & Malthe-Sørenssen, Anders (2020). Using computational essays to scaffold professional physics practice. European Journal of Physics. ISSN 0143-0807. 42(015701). doi: 10.1088/1361-6404/abb8b7. Full text in Research Archive Show summary

  This article describes a curricular innovation designed to help students experience authentic physics inquiry with an emphasis on computational modeling and scientific communication. The educational design centers on a new type of assignment called a computational essay, which was developed and implemented over the course of two semesters of an intermediate electricity and magnetism course at the University of Oslo, Norway. We describe the motivation, learning goals, and scaffolds used in the computational essay project, with the intention that other educators will be able to replicate and adapt our design. We also report on initial findings from this implementation, including key features of student-written computational essays, student reflections on the inquiry process, and self-reported conceptual and attitudinal development. Based on these findings, we argue that computational essays can serve a key role in introducing students to open-ended, inquiry-based work and setting the foundation for future computational research and studies.

• Sveinsson, Henrik Andersen; Hafreager, Anders; Kalia, Rajiv K.; Nakano, Aiichiro; Vashishta, Priya & Malthe-Sørenssen, Anders (2020). Direct Atomic Simulations of Facet Formation and Equilibrium Shapes of SiC Nanoparticles. Crystal Growth & Design. ISSN 1528-7483. 20(4), p. 2147–2152. doi: 10.1021/acs.cgd.9b00612. Full text in Research Archive Show summary

  Understanding the shapes of nanoparticles is an important interdisciplinary problem because particle shapes can affect their properties, functionality, and applications. Advances in nanoscale imaging probes have revealed exquisite details of nanofaceting phenomena. However, quantitative theoretical predictions have not kept up the pace with experimental advances, and the atomic pathways of facet formation are largely unknown due to a lack of direct observations and simulations. Here we examine facet formation in spherical and cubic SiC nanoparticles and in SiC nanowires using molecular dynamics simulations reaching microseconds. We characterize layer-by-layer formation, diffusional motion along edges and corners, and determine energy barriers. We find that the equilibrium shapes are identical regardless of the initial shape of SiC nanoparticles or nanowires. For spherical and cubic nanoparticles, (110) facets form within 10 ns by lateral liquid-like diffusion of atoms. In contrast, faceting in SiC nanowires also involves normal diffusional motion with a higher energy barrier and hence much longer faceting times. These results have important implications for molecular-level understanding of the synthesis and stability of ceramic nanocrystals and nanowires.

• Hansson, Kenth-Arne; Eftestøl, Einar; Bruusgaard, Jo C.; Juvkam, Inga; Cramer, Alyssa W. & Malthe-Sørenssen, Anders [Show all 8 contributors for this article] (2020). Myonuclear content regulates cell size with similar scaling properties in mice and humans. Nature Communications. 11(1). doi: 10.1038/s41467-020-20057-8. Full text in Research Archive Show summary

  Muscle fibers are the largest cells in the body, and one of its few syncytia. Individual cell sizes are variable and adaptable, but what governs cell size has been unclear. We find that muscle fibers are DNA scarce compared to other cells, and that the nuclear number (N) adheres to the relationship N = aVb where V is the cytoplasmic volume. N invariably scales sublinearly to V (b < 1), making larger cells even more DNA scarce. N scales linearly to cell surface in adult humans, in adult and developing mice, and in mice with genetically reduced N, but in the latter the relationship eventually fails when they reach adulthood with extremely large myonuclear domains. Another exception is denervation-atrophy where nuclei are not eliminated. In conclusion, scaling exponents are remarkably similar across species, developmental stages and experimental conditions, suggesting an underlying scaling law where DNA-content functions as a limiter of muscle cell size.

• Gløersen, Øyvind Nøstdahl; Gilgien, Matthias; Dysthe, Dag Kristian; Malthe-Sørenssen, Anders & Losnegard, Thomas Johansen (2020). Oxygen Demand, Uptake, and Deficits in Elite Cross-country Skiers during a 15-km Race. Medicine & Science in Sports & Exercise. ISSN 0195-9131. 52(4), p. 983–992. doi: 10.1249/MSS.0000000000002209. Full text in Research Archive Show summary

  Purpose: To quantify the repeated oxygen deficits attained during intermittent endurance exercise by measuring oxygen consumption (V[Combining Dot Above]O2) and oxygen demand (V[Combining Dot Above]O2dem) throughout a simulated roller ski race. Methods: Eight male elite cross-country skiers (V[Combining Dot Above]O2,peak 77.4 ± 4.4 mL·min-1·kg-1) raced a 13.5 km roller ski time-trial on a World Cup course. On two additional days, athletes completed (i) 6 sub-maximal loads (à 5 min) and a ~ 4 min maximal trial to establish athlete-specific estimates of skiing economy, V[Combining Dot Above]O2,peak, and maximal ΣO2def (MAOD); and (ii) a simulation of the time-trial on a roller skiing treadmill. During the simulation, external work rate (Pprop) and skiing speed (v) were adjusted to match the Pprop and v measured during the time-trial, and pulmonary V[Combining Dot Above]O2 was measured breath-by-breath. V[Combining Dot Above]O2dem and ΣO2def were calculated using an athlete-specific model for skiing economy throughout the treadmill simulation. Results: During the treadmill simulation V[Combining Dot Above]O2 was on average 0.77V[Combining Dot Above]O2,peak, and active V[Combining Dot Above]O2dem, (i.e., excluding the time in simulated downhills), was on average 1.01V[Combining Dot Above]O2,peak. The athletes repeatedly attained substantial oxygen deficits in individual uphill sections of the treadmill simulation, but the deficits were typically small compared to their MAOD (average 14%, range ~0-50%). However, the ΣO2def summed over all periods of active propulsion was on average 3.8MAOD. Conclusion: Athletes repeatedly attain substantial oxygen deficits in the uphill segments of a distance cross-country ski race. Furthermore, the total accumulated oxygen deficit of all these segments is several times higher than the athletes’ MAODs. This suggests that rapid recovery of the energy stores represented by the oxygen deficit is necessary during downhill sections, and that this might be an important determinant of distance skiing performance.

• Lepperød, Mikkel Elle; Dragly, Svenn-Arne; Buccino, Alessio Paolo; Mobarhan, Milad; Malthe-Sørenssen, Anders & Hafting, Torkel [Show all 7 contributors for this article] (2020). Experimental Pipeline (Expipe): A Lightweight Data Management Platform to Simplify the Steps From Experiment to Data Analysis. Frontiers in Neuroinformatics. 14. doi: 10.3389/fninf.2020.00030. Full text in Research Archive Show summary

  As experimental neuroscience is moving toward more integrative approaches, with a variety of acquisition techniques covering multiple spatiotemporal scales, data management is becoming increasingly challenging for neuroscience laboratories. Often, datasets are too large to practically be stored on a laptop or a workstation. The ability to query metadata collections without retrieving complete datasets is therefore critical to efficiently perform new analyses and explore the data. At the same time, new experimental paradigms lead to constantly changing specifications for the metadata to be stored. Despite this, there is currently a serious lack of agile software tools for data management in neuroscience laboratories. To meet this need, we have developed Expipe, a lightweight data management framework that simplifies the steps from experiment to data analysis. Expipe provides the functionality to store and organize experimental data and metadata for easy retrieval in exploration and analysis throughout the experimental pipeline. It is flexible in terms of defining the metadata to store and aims to solve the storage and retrieval challenges of data/metadata due to ever changing experimental pipelines. Due to its simplicity and lightweight design, we envision Expipe as an easy-to-use data management solution for experimental laboratories, that can improve provenance, reproducibility, and sharing of scientific projects.

• Ponthus, Nicolas; Scheibert, Julien; Thøgersen, Kjetil; Malthe-Sørenssen, Anders & Perret-Liaudet, Joel (2019). Statistics of the separation between sliding rigid rough surfaces: Simulations and extreme value theory approach. Physical Review E (PRE). ISSN 2470-0045. 99(2). doi: 10.1103/PhysRevE.99.023004. Full text in Research Archive Show summary

  When a rigid rough solid slides on a rigid rough surface, it experiences a random motion in the direction normal to the average contact plane. Here, through simulations of the separation at single-point contact between self-affine topographies, we characterize the statistical and spectral properties of this normal motion. In particular, its rms amplitude is much smaller than that of the equivalent roughness of the two topographies and depends on the ratio of the slider's lateral size over a characteristic wavelength of the topography. In addition, due to the nonlinearity of the sliding contact process, the normal motion's spectrum contains wavelengths smaller than the smallest wavelength present in the underlying topographies. We show that the statistical properties of the normal motion's amplitude are well captured by a simple analytic model based on the extreme value theory framework, extending its applicability to sliding-contact-related topics.

• Thøgersen, Kjetil; Sveinsson, Henrik Andersen; Scheibert, Julien; Renard, Francois & Malthe-Sørenssen, Anders (2019). The Moment Duration Scaling Relation for Slow Rupture Arises From Transient Rupture Speeds. Geophysical Research Letters. ISSN 0094-8276. 46(22), p. 12805–12814. doi: 10.1029/2019GL084436. Full text in Research Archive Show summary

  Plain Language Summary Observations have shown that the duration of earthquakes is related to the seismic moment through a power law. The power law exponent is different for regular earthquakes and slow aseismic rupture, and the origin of this difference is currently debated in the literature. In this letter, we introduce a minimal mechanical friction model that contains both slow and regular earthquakes and demonstrate that the different power laws emerge naturally within the model because the propagation speed of slow earthquakes decays as a power law in time, whereas the propagation speed of regular earthquakes remains fairly constant.

• Thøgersen, Kjetil; Sveinsson, Henrik Andersen; Amundsen, David Skålid; Scheibert, Julien; Renard, Francois & Malthe-Sørenssen, Anders (2019). Minimal model for slow, sub-Rayleigh, supershear, and unsteady rupture propagation along homogeneously loaded frictional interfaces. Physical Review E (PRE). ISSN 2470-0045. 100(4). doi: 10.1103/PhysRevE.100.043004. Full text in Research Archive Show summary

  In nature and experiments, a large variety of rupture speeds and front modes along frictional interfaces are observed. Here, we introduce a minimal model for the rupture of homogeneously loaded interfaces with velocity strengthening dynamic friction, containing only two dimensionless parameters; ¯ τ , which governs the prestress, and ¯¯¯ α , which is set by the interfacial viscosity. This model contains a large variety of front types, including slow fronts, sub-Rayleigh fronts, supershear fronts, slip pulses, cracks, arresting fronts, and fronts that alternate between arresting and propagating phases. Our results indicate that this wide range of front types is an inherent property of frictional systems with velocity strengthening branches.

• Thøgersen, Kjetil; Gilbert, Adrien; Schuler, Thomas & Malthe-Sørenssen, Anders (2019). Rate-and-state friction explains glacier surge propagation. Nature Communications. 10(1). doi: 10.1038/s41467-019-10506-4. Full text in Research Archive Show summary

  The incomplete understanding of glacier dynamics is a major source of uncertainty in assessments of sea-level rise from land-based ice. Through increased ice discharge into the oceans, accelerating glacier flow has the potential to considerably enhance expected sea-level change, well ahead of scenarios considered by the IPCC. Central in our incomplete understanding is the motion at the glacier bed, responsible for flow transients and instabilities involving switches from slow to fast flow. We introduce a rate-and-state framework for the transient evolution of basal shear stress, which we incorporate in glacier simulations. We demonstrate that a velocity-strengthening-weakening transition combined with a characteristic length scale for the opening of subglacial cavities is sufficient to reproduce several previously unexplained features of glacier surges. The rate-and-state framework opens for new ways to analyze, understand and predict transient glacier dynamics as well as to assess the stability of glaciers and ice caps.

• Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2019). Molecular-scale thermally activated fractures in methane hydrates: a molecular dynamics study. Physical Chemistry, Chemical Physics - PCCP. ISSN 1463-9076. 21(25), p. 13539–13544. doi: 10.1039/c9cp01337g. Full text in Research Archive Show summary

  We perform multiple large molecular dynamics simulations to study the fracture behaviour of monocrystalline methane hydrates under tension. We examine the fracture initiation phase and find that the fracture process can be divided into two phases: slow crack growth and rapid crack propagation. The time of the slow crack growth phase can be predicted by a thermal activation model [L. Vanel et al., J. Phys. D: Appl. Phys., 2009, 42, 214007] where an energy barrier has to be overcome in order for the crack to propagate. Our simulations predict that the slow growth phase vanishes when the stress intensity factor approaches Image ID:c9cp01337g-t1.gif.

• Solbrå, Andreas Våvang; Bergersen, Aslak Wigdahl; Brink, Jonas van den; Malthe-Sørenssen, Anders; Einevoll, Gaute & Halnes, Geir (2018). A Kirchhoff-Nernst-Planck framework for modeling large scale extracellular electrodiffusion surrounding morphologically detailed neurons. PLoS Computational Biology. ISSN 1553-734X. 14(10), p. 1–26. doi: 10.1371/journal.pcbi.1006510. Full text in Research Archive Show summary

  Many pathological conditions, such as seizures, stroke, and spreading depression, are associated with substantial changes in ion concentrations in the extracellular space (ECS) of the brain. An understanding of the mechanisms that govern ECS concentration dynamics may be a prerequisite for understanding such pathologies. To estimate the transport of ions due to electrodiffusive effects, one must keep track of both the ion concentrations and the electric potential simultaneously in the relevant regions of the brain. Although this is currently unfeasible experimentally, it is in principle achievable with computational models based on biophysical principles and constraints. Previous computational models of extracellular ion-concentration dynamics have required extensive computing power, and therefore have been limited to either phenomena on very small spatiotemporal scales (micrometers and milliseconds), or simplified and idealized 1-dimensional (1-D) transport processes on a larger scale. Here, we present the 3-D Kirchhoff-Nernst-Planck (KNP) framework, tailored to explore electrodiffusive effects on large spatiotemporal scales. By assuming electroneutrality, the KNP-framework circumvents charge-relaxation processes on the spatiotemporal scales of nanometers and nanoseconds, and makes it feasible to run simulations on the spatiotemporal scales of millimeters and seconds on a standard desktop computer. In the present work, we use the 3-D KNP framework to simulate the dynamics of ion concentrations and the electrical potential surrounding a morphologically detailed pyramidal cell. In addition to elucidating the single neuron contribution to electrodiffusive effects in the ECS, the simulation demonstrates the efficiency of the 3-D KNP framework. We envision that future applications of the framework to more complex and biologically realistic systems will be useful in exploring pathological conditions associated with large concentration variations in the ECS.

• Dragly, Svenn-Arne; Mobarhan, Milad; Lepperød, Mikkel Elle; Tennøe, Simen; Fyhn, Marianne & Hafting, Torkel [Show all 7 contributors for this article] (2018). Experimental directory structure (Exdir): An alternative to HDF5 without introducing a new file format. Frontiers in Neuroinformatics. 12. doi: 10.3389/fninf.2018.00016. Full text in Research Archive
• Aupart, Claire Olga Maryse; Dunkel, Kristina G; Angheluta, Luiza; Austrheim, Håkon Olaf; Ildefonse, Beniot & Malthe-Sørenssen, Anders [Show all 7 contributors for this article] (2018). Olivine Grain Size Distributions in Faults and Shear Zones: Evidence for Nonsteady State Deformation. Journal of Geophysical Research (JGR): Solid Earth. ISSN 2169-9313. 123(9), p. 7421–7443. doi: 10.1029/2018JB015836. Full text in Research Archive Show summary

  The grain size distribution of deformed rocks may provide valuable information about their deformation history and the associated mechanisms. Here we present a unique set of olivine grain size distributions from ultramafic rocks deformed under a wide range of stress and strain rate conditions. Both experimentally deformed and naturally deformed samples are included. We observe a surprisingly uniform behavior, and most samples show power law grain size distributions. Convincing lognormal distributions across all scales were only observed for samples experimentally deformed at high temperature (1200 °C) and for some mantle‐deformed natural samples. Single power law distributions were observed for natural samples deformed by brittle mechanisms and by samples deformed experimentally in the regime of low‐temperature plasticity. Most natural samples show a crossover in power law scaling behavior near the median grain size from a steep slope for the larger grain fraction to a more gentle slope for the smaller grains. The small grain fraction shows a good data collapse when normalized to the crossover length scale. The associated power law slope indicates a common grain size controlling process. We propose a model that explains how such a scaling behavior may arise in the dislocation creep regime from the competition between the rate involved in the dislocation dynamics and the imposed strain rate. The common departure from lognormal distributions suggests that naturally deformed samples often have a deformation history that is far from a steady state scenario and probably reflects deformation under highly variable stress and strain rates. ©2018. American Geophysical Union. All Rights Reserved.

• Thompson, Elise Holter; Lensjø, Kristian Kinden; Wigestrand, Mattis B; Malthe-sørenssen, Anders; Hafting, Torkel & Fyhn, Marianne (2018). Removal of perineuronal nets disrupts recall of a remote fear memory. Proceedings of the National Academy of Sciences of the United States of America (PNAS). ISSN 0027-8424. 115(3), p. 607–612. doi: 10.1073/pnas.1713530115. Full text in Research Archive Show summary

  Throughout life animals learn to recognize cues that signal danger and instantaneously initiate an adequate threat response. Memories of such associations may last a lifetime and far outlast the intracellular molecules currently found to be important for memory processing. The memory engram may be supported by other more stable molecular components, such as the extracellular matrix structure of perineuronal nets (PNNs). Here, we show that recall of remote, but not recent, visual fear memories in rats depend on intact PNNs in the secondary visual cortex (V2L). Supporting our behavioral findings, increased synchronized theta oscillations between V2L and basolateral amygdala, a physiological correlate of successful recall, was absent in rats with degraded PNNs in V2L. Together, our findings suggest a role for PNNs in remote memory processing by stabilizing the neural network of the engram.

• Gløersen, Øyvind Nøstdahl; Losnegard, Thomas Johansen; Malthe-Sørenssen, Anders; Dysthe, Dag Kristian & Gilgien, Matthias (2018). Propulsive Power in Cross-Country Skiing: Application and Limitations of a Novel Wearable Sensor-Based Method During Roller Skiing. Frontiers in Physiology. 9, p. 1–14. doi: 10.3389/fphys.2018.01631. Full text in Research Archive Show summary

  Cross-country skiing is an endurance sport that requires extremely high maximal aerobic power. Due to downhill sections where the athletes can recover, skiersmust also have the ability to perform repeated efforts where metabolic power substantially exceeds maximal aerobic power. Since the duration of these supra-aerobic efforts is often in the order of seconds, heart rate, and pulmonary VO2 do not adequately reflect instantaneous metabolic power. Propulsive power (Pprop) is an alternative parameter that can be used to estimate metabolic power, but the validity of such calculations during cross-country skiing has rarely been addressed. The aim of this study was therefore twofold: to develop a procedure using small non-intrusive sensors attached to the athlete for estimating Pprop during roller-skiing and to evaluate its limits; and (2) to utilize this procedure to determine the Pprop generated by high-level skiers during a simulated distance race. Eight elite male cross-country skiers simulated a 15 km individual distance race on roller skis using ski skating techniques on a course (13.5 km) similar to World Cup skiing courses. Pprop was calculated using a combination of standalone and differential GNSS measurements and inertial measurement units. The method’s measurement error was assessed using a Monte Carlo simulation, sampling from the most relevant sources of error. Pprop decreased approximately linearly with skiing speed and acceleration, and was approximated by the equation Pprop(v, ˙v) = −0.54·v −0.71·˙v + 7.26 W·kg−1. Pprop was typically zero for skiing speeds >9 m·s−1, because the athletes transitioned to the tuck position. Peak Pprop was 8.35 ± 0.63 W·kg−1 and was typically attained during the final lap in the last major ascent, while average Pprop throughout the race was 3.35 ± 0.23 W·kg−1. The measurement error of Pprop increased with skiing speed, from 0.09 W·kg−1 at 2.0 m·s−1 to 0.58 W·kg−1 at 9.0 m·s−1. In summary, this study is the first to provide continuous measurements of Pprop for distance skiing, as well as the first to quantify the measurement error during roller skiing using the power balance principle. Therefore, these results provide novel insight into the pacing strategies employed by high-level skiers.

• Plümper, Oliver; Botan, Alexandru; Los, Catharina; Liu, Yang; Malthe-Sørenssen, Anders & Jamtveit, Bjørn (2017). Fluid-driven metamorphism of the continental crust governed by nanoscale fluid flow. Nature Geoscience. ISSN 1752-0894. 10(9), p. 685–690. doi: 10.1038/ngeo3009. Full text in Research Archive Show summary

  The transport of fluids through the Earth’s crust controls the redistribution of elements to form mineral and hydrocarbon deposits, the release and sequestration of greenhouse gases, and facilitates metamorphic reactions that influence lithospheric rheology. In permeable systems with a well-connected porosity, fluid transport is largely driven by fluid pressure gradients. In less permeable rocks, deformation may induce permeability by creating interconnected heterogeneities, but without these perturbations, mass transport is limited along grain boundaries or relies on transformation processes that self-generate transient fluid pathways. The latter can facilitate large-scale fluid and mass transport in nominally impermeable rocks without large-scale fluid transport pathways. Here, we show that pervasive, fluid-driven metamorphism of crustal igneous rocks is directly coupled to the production of nanoscale porosity. Using multi-dimensional nano-imaging and molecular dynamics simulations, we demonstrate that in feldspar, the most abundant mineral family in the Earth’s crust, electrokinetic transport through reaction-induced nanopores (<100 nm) can potentially be significant. This suggests that metamorphic fluid flow and fluid-mediated mineral transformation reactions can be considerably influenced by nanofluidic transport phenomena.

• Dragly, Svenn-Arne; Mobarhan, Milad; Solbrå, Andreas Våvang; Tennøe, Simen; Hafreager, Anders & Malthe-Sørenssen, Anders [Show all 9 contributors for this article] (2017). Neuronify: An educational simulator for neural circuits. eNeuro. 4(2:e0022-17.2017), p. 1–13. doi: 10.1523/ENEURO.0022-17.2017. Full text in Research Archive Show summary

  Educational software (apps) can improve science education by providing an interactive way of learning about complicated topics that are hard to explain with text and static illustrations. However, few educational apps are available for simulation of neural networks. Here, we describe an educational app, Neuronify, allowing the user to easily create and explore neural networks in a plug-and-play simulation environment. The user can pick network elements with adjustable parameters from a menu, i.e., synaptically connected neurons modelled as integrate-and-fire neurons and various stimulators (current sources, spike generators, visual, and touch) and recording devices (voltmeter, spike detector, and loudspeaker). We aim to provide a low entry point to simulation-based neuroscience by allowing students with no programming experience to create and simulate neural networks. To facilitate the use of Neuronify in teaching, a set of premade common network motifs is provided, performing functions such as input summation, gain control by inhibition, and detection of direction of stimulus movement. Neuronify is developed in C++ and QML using the cross-platform application framework Qt and runs on smart phones (Android, iOS) and tablet computers as well personal computers (Windows, Mac, Linux).

• Thøgersen, Kjetil; Dabrowski, Marcin & Malthe-Sørenssen, Anders (2016). Transient cluster formation in sheared non-Brownian suspensions. Physical Review E (PRE). ISSN 2470-0045. 93(2). doi: 10.1103/PhysRevE.93.022611. Full text in Research Archive
• Huang, Hai; Meakin, Paul & Malthe-Sørenssen, Anders (2016). Physics-based simulation of multiple interacting crack growth in brittle rocks driven by thermal cooling. International journal for numerical and analytical methods in geomechanics (Print). ISSN 0363-9061. 40(16), p. 2163–2177. doi: 10.1002/nag.2523. Full text in Research Archive
• Trømborg, Jørgen Kjoshagen; Sveinsson, Henrik Andersen; Thøgersen, Kjetil; Scheibert, Julien & Malthe-Sørenssen, Anders (2015). Speed of fast and slow rupture fronts along frictional interfaces. Physical Review E. Statistical, Nonlinear, and Soft Matter Physics. ISSN 1539-3755. 92(1). doi: 10.1103/PhysRevE.92.012408. Full text in Research Archive Show summary

  The transition from stick to slip at a dry frictional interface occurs through the breaking of microjunctions between the two contacting surfaces. Typically, interactions between junctions through the bulk lead to rupture fronts propagating from weak and/or highly stressed regions, whose junctions break first. Experiments find rupture fronts ranging from quasistatic fronts, via fronts much slower than elastic wave speeds, to fronts faster than the shear wave speed. The mechanisms behind and selection between these fronts are still imperfectly understood. Here we perform simulations in an elastic two-dimensional spring-block model where the frictional interaction between each interfacial block and the substrate arises from a set of junctions modeled explicitly. We find that material slip speed and rupture front speed are proportional across the full range of front speeds we observe. We revisit a mechanism for slow slip in the model and demonstrate that fast slip and fast fronts have a different, inertial origin. We highlight the long transients in front speed even along homogeneous interfaces, and we study how both the local shear to normal stress ratio and the local strength are involved in the selection of front type and front speed. Last, we introduce an experimentally accessible integrated measure of block slip history, the Gini coefficient, and demonstrate that in the model it is a good predictor of the history-dependent local static friction coefficient of the interface. These results will contribute both to building a physically based classification of the various types of fronts and to identifying the important mechanisms involved in the selection of their propagation speed.

View all works in NVA

• Reinås, Karl Henrik Storhaug; Pedersen, Tuva Emilie Orheim; Aasland, Sigrun; Malthe-Sørenssen, Anders; Grande, Trine Skei & Blikstad-Balas, Marte [Show all 10 contributors for this article] (2026). Hvor er pensum? doi: https:/www.youtube.com/watch?v=exd7Z9267II. Full text in Research Archive Show summary

  Hvor får vi kunnskapen fra i 2026 og hvordan? Fra YouTube, i en app, i en bok, via KI – eller helt andre steder? Må man lese pensum for å bestå eksamen, og må man lese mange bøker for å ta doktorgrad? Hvor er pensum? Må man gå på forelesninger, og hvordan skal studenter lære? Må man være god til å lese for å lære? Hvordan påvirker kunstig intelligens læring, og er det mye hjelp å få i sammendragene? Forleggerforeningen rapporterte i 2024 at norske studenter i snitt kjøper 1,9 nye pensumbøker i året. Salget av lærebøker er i fritt fall. Er boka blitt overflødig? Program: Introduksjon/utspørring: Hvor finner dere kunnskapen? Hvor mye leser dere? Trenger man å lese pensum? Holder det med sammendrag? Karl Henrik Reinås, leder av Stipendiatorganisasjonene i Norge Tuva Orheim Pedersen, masterstudent ved Oslomet Paneldebatt: Sigrun Aasland, forsknings- og høyere utdanningsminister Anders Malthe-Sørenssen, professor og leder av KI-utvalget, UiO Trine Skei Grande, direktør i Forleggerforeningen Marte Blikstad-Balas, professor og leseforsker ved Institutt for lærerutdanning og skoleforskning, UiO Duell: Kan en læringsapp erstatte pensum? Sykepleierforbundet samarbeider med læringsappen MedEasy i pensumdrilling av studentene. De lover "100 prosent faglig innhold blandet med rare ablegøyer". Og studentene møter opp mot betaling. Henriette Karlsen Brenna, leder i Sykepleierforbundet student Marit Kirkevold, dekan ved Fakultet for helsevitenskap ved OsloMet Debattledere er Tove Lie, redaktør i Khrono, og Hilde Kristin Strand, journalist i Khrono. Arrangementet er i regi av Kunnskapsavisen Khrono og Forleggerforeningen.

• Malthe-Sørenssen, Anders; Harang, Fabian Andsem; Johnson, Victoria Christine Årang; Nyrup, Jacob; Lie, Knut-Andreas & Sandve, Geir Kjetil Ferkingstad [Show all 7 contributors for this article] (2026). Hvordan endrer KI-agenter forskningsprosessen? Full text in Research Archive Show summary

  Dette møtet vil ta for seg de forskningsetiske spørsmålene som reiser seg i møte med bruken av KI-agenter i forskning

• Goodwin, Morten; Mørch, Hedda Hassel; Malthe-Sørenssen, Anders & Inntjore, Hilde (2025). Panelsamtale: Hva gjør KI med høyere utdanning, og hva gjør høyere utdanning med KI? Full text in Research Archive
• Camposano, Anthony Val Canillas; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2024). Machine Learning an All-Atom Empirical Force Field for a Water-Silica System. Full text in Research Archive Show summary

  Molecular modeling of the interaction between silica and water provides a better understanding of how water affects rocks and glasses at the nanoscale. Here, we parametrize a new classical dissociative force field for the water–silica system. We create a classical force field capable of reactive interactions between all chemical species in the water–silica system, using the same functional form by Vashishta et al. to describe bulk water, silica, and silica–water interactions. Using a hierarchical genetic algorithm as a global optimizer and the Nelder-Mead simplex algorithm as a local optimizer, we parameterized the Vashishta force field to reproduce the density, transport properties, and vapor-liquid properties of water. We then used the same method to obtain water–silica interaction parameters to reproduce gas phase orthosilicic geometry and silica surface properties such as silanol concentration, heat of immersion, and the wettability angle of water on a silica surface. For oxygen–silicon interactions in silica, we reuse an existing Vashishta potential parameter set that has been parameterized to reproduce bulk mechanical behavior, surface properties, and fracture properties. Our developed force field can be used to study processes such as dissolution, friction, and stress corrosion fracture in the presence of water over a wide range of thermodynamic conditions for larger-scale and longer-scale simulations.

• Tarigopula, Homa; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2023). Model-based meta-reinforcement learning for alchemy. Full text in Research Archive Show summary

  Meta-learning emphasises the ‘learning to learn’ paradigm where the similar characteristics between the tasks generate a common basis bearing relevance that extends beyond the trained tasks and acts as an inductive bias that helps the learning agent to quickly generalize to unseen tasks from the same distribution. For example during navigation, understanding the underlying structure of the environment can aid inferences when going back to an explored state from an unexplored point of view. Recent work by Whittington et al 2020 has shown that the underlying graph defining the structure once learnt is reusable across different environments consisting of different sensory observations, thus forming an invariant feature basis. This can be extended to the relational inference scenarios encountered by reinforcement learning (RL) agents. A recently introduced benchmark ‘Alchemy' aims at challenging meta-RL by providing both a convenient interface and an optimal solution for assessing agent efficiency. The meta-RL agent navigating the alchemy environment is challenged with understanding the latent causal structure to excel at this task with maximum reward. We aim to train model-based meta-RL agents to solve alchemy, by introducing useful inductive bias for capturing the underlying structure of explored environments. We hypothesize that training such an agent, in addition to improving the meta-RL knowledge base, may help shed light on the interactions between Prefrontal cortex and Entorhinal cortex. Here, we present preliminary results from initial studies that show promising performance on the task compared to the two deep RL agents (VMPO, IMPALA) previously deployed to solve Alchemy.

• Najafi, Fahimeh; Sveinsson, Henrik Andersen; Dreierstad, Christer & Malthe-Sørenssen, Anders (2023). Modeling the relationship between mechanical yield stress and material geometry using convolutional neural networks. doi: 10.5194/egusphere-egu23-13293. Full text in Research Archive Show summary

  Neural networks have proved to be able to capture the relevant and informative features of a wide variety of data types and predict the desired output for different regression or classification problems. Finding a mapping between materials&#8217; structure and a given physical property of those systems is an example of a problem that could be approached with machine learning methods like neural networks. Especially when we are dealing with systems with a very large design space where using classical computational methods like molecular dynamics can be very time and resource-consuming for the study of a very large number of systems, a well-trained neural network can be greatly faster and more efficient for computing the relevant properties. In this work, we study &#945;-quartz crystals with one porous layer with simplex noise as the shape of porosity. Simplex noise is a gradient based procedural algorithm that can produce irregular geometries with surface morphology resembling what is observed in nature. The property that we want the neural network to learn is the yield stress of these systems under both shear and tensile deformation. Molecular dynamics simulations are used for a randomly selected sample of possible structures in order to generate the ground truth to be used as the training data. We employ deep convolutional neural networks (CNN) which are commonly used when dealing with image or image-like data since the input data for the problem in hand is a binary 2-D structure of the porous layer of the systems. The trained model is compared with a basic polynomial fit of stress versus porosity. The trained CNN is successful in predicting the yield stress of a system based on the geometry of that given system, with lower variability and higher precision compared to the base polynomial regression method. The saliency maps created with the trained model show the model to be successful in capturing the physics of the problem when compared with the stress fields calculated using molecular dynamics simulations. This method of modeling materials can be further developed for the inverse design of structures with desired properties without the need for a huge number of simulations on a wide domain of possible systems.

• Bakke, Jostein; Langford, Malcolm; Børrud, Elin; Cotner, Sehoya Harris; Engebretsen, Eivind & Graver, Fredrik (Georg Fredrik) [Show all 14 contributors for this article] (2023). Fremtidens høyere utdanning trenger en oppgradering - ikke budsjettkutt! Full text in Research Archive
• Skogvoll, Vidar; Malthe-Sørenssen, Anders & Gregers, Tone Fredsvik (2023). Real læring - ChatGPT: Roboten som vil endre læring og undervisning. [Radio]. Podcast. Full text in Research Archive
• Sveinsson, Henrik Andersen; Thøgersen, Kjetil & Malthe-Sørenssen, Anders (2023). Grain boundary behavior of hydrate–hydrate and ice—ice bicrystals: molecular dynamics insights. Full text in Research Archive
• Camposano, Anthony Val Canillas; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2023). Tuning Vashishta Potential for Water. Full text in Research Archive
• Holzhausen, Konstantin; Lepperød, Mikkel & Malthe-Sørenssen, Anders (2022). Robustness of neural networks trained with biologically inspired local learning rules. Full text in Research Archive
• Schøyen, Vemund; Pettersen, Markus; Holzhausen, Konstantin; Malthe-Sørenssen, Anders & Lepperød, Mikkel (2022). Navigating Multiple Environments with Emergent Grid Cell Remapping. Full text in Research Archive
• Schøyen, Vemund; Pettersen, Markus; Holzhausen, Konstantin; Malthe-Sørenssen, Anders & Lepperød, Mikkel (2022). Place and grid cell navigation in multiple environments. Full text in Research Archive
• Rogge, Frederik; Malthe-Sørenssen, Anders; Fyhn, Marianne & Lepperød, Mikkel (2022). On the Structural Sensitivity of the Successor Representation. Full text in Research Archive
• Pettersen, Markus; Schøyen, Vemund; Malthe-Sørenssen, Anders & Lepperød, Mikkel (2022). Emergent Place Cells and Remapping in a Recurrent Neural Network Model. Full text in Research Archive
• Targopula, Homa; Malthe-Sørenssen, Anders & Lepperød, Mikkel (2022). Model-based meta reinforcement learning for alchemy. Full text in Research Archive
• Tarigopula, Homa; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2022). Model based meta reinforcement learning for alchemy. doi: https:/fensforum.org/wp-content/uploads/2023/01/FENS2022.pdf. Full text in Research Archive Show summary

  Meta-learning emphasises the ‘learning to learn’ paradigm where the similar characteristics between the tasks generate a common basis bearing relevance that extends beyond the trained tasks and acts as an inductive bias that helps the learning agent to quickly generalize to unseen tasks from the same distribution. For example during navigation, understanding the underlying structure of the environment can aid inferences when going back to an explored state from an unexplored point of view. Recent work by Whittington et al 2020 has shown that the underlying graph defining the structure once learnt is reusable across different environments consisting of different sensory observations, thus forming an invariant feature basis. This can be extended to the relational inference scenarios encountered by reinforcement learning (RL) agents. A recently introduced benchmark ‘Alchemy' aims at challenging meta-RL by providing both a convenient interface and an optimal solution for assessing agent efficiency. The meta-RL agent navigating the alchemy environment is challenged with understanding the latent causal structure to excel at this task with maximum reward. We aim to train model-based meta-RL agents to solve alchemy, by introducing useful inductive bias for capturing the underlying structure of explored environments. We hypothesize that training such an agent, in addition to improving the meta-RL knowledge base, may help shed light on the interactions between Prefrontal cortex and Entorhinal cortex. Here, we present preliminary results from initial studies that show promising performance on the task compared to the two deep RL agents (VMPO, IMPALA) previously deployed to solve Alchemy.

• Schøyen, Vemund Sigmundson; Pettersen, Markus Borud; Holzhousen, Konstantin; Fyhn, Marianne; Malthe-Sørenssen, Anders & Lepperød, Mikkel Elle (2022). Coherently Remapping Toroidal Cells But Not Grid Cells are Responsible for Path Integration in Virtual Agents. BioRxiv. ISSN 0362-4331. doi: 10.1101/2022.08.18.504379. Full text in Research Archive
• Idsøe, Ella Maria Cosmovici & Malthe-Sørenssen, Anders (2021). Experiences from Establishing a Cross-Disciplinary Honors Program at the University of Oslo. Full text in Research Archive
• Maasø, Arnt; Brevik, Lisbeth M.; Vasbø, Kristin Beate; Breivik, Jarle; Lerheim, Birgitte & Malthe-Sørenssen, Anders [Show all 7 contributors for this article] (2021). Dette må bli læringsmiljøets år. Full text in Research Archive
• Sveinsson, Henrik Andersen; Thøgersen, Kjetil & Malthe-Sørenssen, Anders (2020). Grain boundary behavior of hydrate–hydrate and ice–ice bicrystals: Molecular dynamics insights. Full text in Research Archive
• Fyhn, Marianne; Malthe-Sørenssen, Anders; Mørken, Knut Martin; Nederbragt, Alexander Johan; Odden, Tor Ole & Sundnes, Joakim (2019). Beregningsorientering og utdanningsutvikling. Full text in Research Archive
• Odden, Tor Ole & Malthe-Sørenssen, Anders (2019). Computational Essays and Computational Literacy at the University of Oslo. Full text in Research Archive
• Guren, Marthe Grønlie; Zheng, Xiaojiao; Jamtveit, Bjørn; Hafreager, Anders; Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders [Show all 7 contributors for this article] (2019). Molecular dynamics simulations of the hydration force and transport properties of a water film during reaction-induced fracturing. Full text in Research Archive
• Razbani, Mohammad Amin; Zehner, Jennifer Simone; Jettestuen, Espen; Røyne, Anja; Sikorski, Pawel & Malthe-Sørenssen, Anders (2019). A pore-scale model of microbially induced calcium carbonate precipitation. Full text in Research Archive
• Plümper, Oliver; Botan, Alexandru; Los, Catharina; Liu, Yang; Malthe-Sørenssen, Anders & Jamtveit, Bjørn (2019). Erratum to: Fluid-driven metamorphism of the continental crust governed by nanoscale fluid flow (Nature Geoscience, (2017), 10, 9, (685-690), 10.1038/ngeo3009). Nature Geoscience. ISSN 1752-0894. 12(1), p. 80–80. doi: 10.1038/s41561-018-0276-4. Full text in Research Archive
• Thøgersen, Kjetil; Schuler, Thomas & Malthe-Sørenssen, Anders (2019). Nå er isbreene fravristet noen av sine dypeste hemmeligheter. [Journal]. titan.uio.no og Teknisk Ukeblad. Full text in Research Archive
• Thøgersen, Kjetil; Malthe-Sørenssen, Anders; Sveinsson, Henrik Andersen & Scheibert, Julien (2018). Modeling the onset of frictional sliding. Full text in Research Archive
• Solbrå, Andreas Våvang; Bergersen, Aslak Wigdahl; Brink, Jonas van den; Malthe-Sørenssen, Anders; Einevoll, Gaute & Halnes, Geir (2018). Modeling electrodiffusion in the extracellular space around a morphologically detailed neuron. Full text in Research Archive
• Sveinsson, Henrik Andersen; Malthe-Sørenssen, Anders & Ning, Fulong (2018). Mechanisms of shear fracture in polycrystalline methane hydrates. Full text in Research Archive
• Hafreager, Anders; Dragly, Svenn-Arne & Malthe-Sørenssen, Anders (2018). Increased productivity with real time atomistic simulations using Atomify. Full text in Research Archive
• Sveinsson, Henrik Andersen; Hafreager, Anders; Vashishta, Priya; Malthe-Sørenssen, Anders & Kalia, Rajiv K. (2018). Molecular dynamics simulations of faceting of silicon carbide nanoparticles. Full text in Research Archive
• Dragly, Svenn-Arne; Mobarhan, Milad; Lepperød, Mikkel Elle; Tennøe, Simen; Stasik, Alexander Johannes & Fyhn, Marianne [Show all 8 contributors for this article] (2018). Exdir An alternative to HDF5 that uses Numpy files for data, YAML for metadata and simple directories to define the hierarchy. Full text in Research Archive
• Karlsen, Emil; Zehner, Jennifer Simone; Razbani, Mohammad Amin; Jettestuen, Espen; Malthe-sørenssen, Anders & Wentzel, Alexander [Show all 9 contributors for this article] (2018). Metabolic modelling, reactive transport modelling, and experimental verification of microbial-induced processes in the production of bio-concrete. Full text in Research Archive
• Aupart, Claire Olga Maryse; Dunkel, Kristina Grete; Angheluta-Bauer, Luiza; Austrheim, Håkon; Ildefonse, Benoit & Malthe-Sørenssen, Anders [Show all 7 contributors for this article] (2017). Grain size distribution evolution in shear zones: a study case in the Leka Ophiolite Complex (LOC). Full text in Research Archive
• Malthe-sørenssen, Anders; Henriksen, Ellen Karoline; Hjorth-Jensen, Morten; Mørken, Knut Martin; Sølna, Hanne & Tellefsen, Cathrine Wahlstrøm (2017). Center for computing in science education: fornyelse av utdanning ved integrasjon av beregninger. Nordic Journal of STEM Education. 1(1), p. 236–241. Full text in Research Archive
• Malthe-sørenssen, Anders (2017). Center for Computing in Science Education. Full text in Research Archive
• Malthe-sørenssen, Anders (2017). Integration of computing in physics education. Full text in Research Archive
• Malthe-sørenssen, Anders; Sveinsson, Henrik Andersen; Trømborg, Jørgen Kjoshagen; Thøgersen, Kjetil & scheiber, julien (2017). Frictional behavior across scale. Full text in Research Archive
• Malthe-sørenssen, Anders (2017). Integrating computing in undergraduate physics education. Full text in Research Archive
• Malthe-sørenssen, Anders (2017). Dyp digital kompetanse - hvordan sikrer vi at studentene blir kompetente i et internasjonal, digital arbeidsmarked. Full text in Research Archive
• Malthe-sørenssen, Anders (2017). En forskningsnær utdanning ved integrasjon av databeregninger og programmering. Full text in Research Archive
• Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2017). Thermally Activated Slow Crack Evolution in Single Crystal Methane Hydrates. Full text in Research Archive
• Solbrå, Andreas Våvang; Malthe-Sørenssen, Anders; Einevoll, Gaute & Halnes, Geir (2016). A novel electrodiffusive scheme for modeling ion dynamics in neural tissue. Full text in Research Archive
• Solbrå, Andreas Våvang; Einevoll, Gaute; Malthe-Sørenssen, Anders & Halnes, Geir (2016). A novel electrodiffusive scheme for modeling ion dynamics in neural tissue. Full text in Research Archive
• Mørken, Knut Martin; Hjorth-Jensen, Morten; Langtangen, Hans Petter & Malthe-Sørenssen, Anders (2016). Computing across the curriculum. Full text in Research Archive
• Mørken, Knut Martin; Hjorth-Jensen, Morten; Langtangen, Hans Petter & Malthe-Sørenssen, Anders (2016). How should computers influence the mathematics curriculum? Full text in Research Archive
• Ulven, Ole Ivar & Malthe-Sørenssen, Anders (2016). Reaction induced fracturing: chemical-mechanical feedback. Full text in Research Archive
• Sveinsson, Henrik Andersen & Malthe-Sørenssen, Anders (2016). Molecular dynamics modeling of tensile cracks in methane hydrates. Full text in Research Archive Show summary

  Production of methane from hydrates is significantly different from producing from a porous matrix, as the material itself disappears during production. Stress heterogeneities arising when methane hydrates dissociate may trigger fractures in the material, and drive further dissociation through fracture of the hydrate. The unstable nature of the gas hydrates renders it di cult to study its mechanical and failure properties experimentally. Even simple characteristics such as the mode I fracture toughness are hard to obtain. In this study, we use molecular dynamics simulations to obtain fracture properties of methane hydrates, with a main focus on crack initiation. We find that on the nanosecond scale, an important crack initiation mechanism is the slow growth of initial cracks due to melting. Rapid cracks start propagating when the system configuration reach a state close to the Griffith fracture criterion from continuum fracture mechanics. The nature of the process taking the system from an uncritical to a critical state is well described by a relation based on Arrhenius’ equation.

• Mørken, Knut Martin; Hjorth-Jensen, Morten; Langtangen, Hans Petter & Malthe-Sørenssen, Anders (2015). Tanker om digital realfagsundervisning. Full text in Research Archive
• Schöb, Franziska Jessica; Dysthe, Dag Kristian; Scholz, Hanne; Binder, Alexander & Malthe-Sørenssen, Anders (2026). From Pixels to Phenotypes: Deep Learning on Label-Free Microscopy Data for Cell Fate Prediction How Algorithms Help Us See More. Universitetet i Oslo. Full text in Research Archive Show summary

  In this thesis, we used supervised deep learning on label-free microscopy data to predict cell fate and state from purely visual information, without relying on fluorescent labeling or manual segmentation. Across three studies, we combined phase-contrast and holotomographic imaging with convolutional neural networks to tackle problems that are central to stem cell biology and cancer research: predicting the differentiation efficiency of induced pluripotent stem cell clones, identifying germ-layer lineages in differentiating cultures, and detecting macrophage-induced cancer cell death. By building segmentation-free pipelines that operated directly on image patches, we reduced the need for explicit cell masks and made our methods easier to deploy in realistic, noisy experimental settings. The overarching results showed that biological label-free images contained sufficient morphological information for deep networks to recover biologically meaningful patterns, but also revealed clear limits to what static morphology could predict. In stem cell differentiation, our models already distinguished high- from lowefficiency clones early in the protocol, and learned image cues that matched intuitive notions of colony coverage and uniformity. In germlayer prediction, endpoint-trained models decoded lineage identity from holotomographic morphology and reconstructed a plausible timeline of lineage divergence when applied retrospectively, while struggling more in mixed cultures with overlapping lineages. In the cancer cell death study, deep learning distinguished dying from alive cells at the time of death, yet early prediction beyond a short pre-death window remained difficult, indicating that death morphology emerged abruptly rather than gradually over many hours. Together, these findings support the view that deep learning on label-free microscopy could provide non-destructive, scalable readouts of cell fate, while emphasizing the need for time-aware models, richer training labels, and improved domain generalization to fully exploit biological label-free imaging in biomedical workflows.

• Rogge, Frederik Sebastian; Fyhn, Marianne; Malthe-Sørenssen, Anders; Lepperød, Mikkel Elle & Lensjø, Kristian (2026). Normative and Experimental Analyses of Learning-Dependent Representations in the Hippocampal-Parahippocampal System. Universitetet i Oslo. Full text in Research Archive Show summary

  The brain is thought to build a cognitive map of both physical and abstract spaces, helping us navigate the world, organize knowledge, and form new memories. In this thesis, I investigated key brain subregions to uncover the principles underlying the emergence of neural representations. I focused on the general properties of the neural code, how visual inputs are integrated, and how these representations reorganize through learning. To address these questions, I combined large-scale chronic two-photon calcium imaging in mice with normative modeling in artificial neural networks. We show that neurons in the medial entorhinal cortex (MEC) develop task-relevant tuning during learning of a purely visual association task, pointing to a broader role beyond its classical link to spatial coding. In parallel, selective disruption of specific neuron types in the postrhinal cortex (POR) impaired learning of the same task, demonstrating their causal contribution. Furthermore, using artificial neural networks, we uncovered computational principles governing how spatial representations emerge and transform. Finally, we revealed dissociable contributions of POR and MEC to visually guided navigation, highlighting their distinct roles. Together, these results shed new light on the emergence of neural representations in different brain regions and their transformation through learning.

View all works in NVA