arXiv subjects

Christian Rummel

Publications and source records attributed to Christian Rummel.

12 recordsLinked to original sources

arXiv2025

Exploring Robustness of Cortical Morphometry in the presence of white matter lesions, using Diffusion Models for Lesion Filling

Cortical thickness measurements from magnetic resonance imaging, an important biomarker in many neurodegenerative and neurological disorders, are derived by many tools from an initial voxel-wise tissue segmentation. White matter (WM) hypointensities in T1-weighted imaging, such as those arising from multiple sclerosis or small vessel disease, are known to affect the output of brain segmentation methods and therefore bias cortical thickness measurements. These effects are well-documented among traditional brain segmentation tools but have not been studied extensively in tools based on deep-learning segmentations, which promise to be more robust. In this paper, we explore the potential of deep learning to enhance the accuracy and efficiency of cortical thickness measurement in the presence of WM lesions, using a high-quality lesion filling algorithm leveraging denoising diffusion networks. A pseudo-3D U-Net architecture trained on the OASIS dataset to generate synthetic healthy tissue, conditioned on binary lesion masks derived from the MSSEG dataset, allows realistic removal of white matter lesions in multiple sclerosis patients. By applying morphometry methods to patient images before and after lesion filling, we analysed robustness of global and regional cortical thickness measurements in the presence of white matter lesions. Methods based on a deep learning-based segmentation of the brain (Fastsurfer, DL+DiReCT, ANTsPyNet) exhibited greater robustness than those using classical segmentation methods (Freesurfer, ANTs).

eess.IV↗

arXiv2023

Multiscale cortical morphometry reveals pronounced regional and scale-dependent variations across the lifespan

Motivation: Characterising the changes in cortical morphology across the lifespan is fundamental for a range of research and clinical applications. Most studies to date have found a monotonic decrease in commonly used morphometrics, such as cortical thickness and volume, across the entire brain with increasing age. Any regional variations reported are subtle changes in the rate of decrease. However, these descriptions of morphological changes have been limited to a single length scale. Here, we delineate the morphological changes associated with the healthy lifespan in multiscale morphometrics. Methods: We applied multiscale morphometric analysis to structural MRI from subjects aged 6-88 years from NKI (n=833) and CamCAN (n=641). These multiscale morphometrics were obtained at both the cortical hemisphere and lobe level. Results: On the level of whole cortical hemispheres, lifespan trajectories show diverging and even opposing trends at different spatial scales, in contrast to the monotonic decreases of volume and thickness described so far. Importantly, larger scales displayed most dramatic changes across the lifespan (up to 60%). More pronounced lobal differences in lifespan trajectories also became apparent in scales over 0.7mm. In a proof-of-principle application in brain age prediction, we also demonstrate added information contributed by multiscale morphometrics. Conclusion: Our study provides a comprehensive multiscale description of lifespan effects on cortical morphology in an age range from 6-88~years. In future, this can form the foundations for a normative model to compare individuals or cohorts, hence identifying multiscale morphological abnormalities. Our results reveal the complementary information contained in different spatial scales, suggesting that morphometrics should not be considered on a single scale, but as functions of length scale.

q-bio.NC↗

arXiv2023

CortexMorph: fast cortical thickness estimation via diffeomorphic registration using VoxelMorph

The thickness of the cortical band is linked to various neurological and psychiatric conditions, and is often estimated through surface-based methods such as Freesurfer in MRI studies. The DiReCT method, which calculates cortical thickness using a diffeomorphic deformation of the gray-white matter interface towards the pial surface, offers an alternative to surface-based methods. Recent studies using a synthetic cortical thickness phantom have demonstrated that the combination of DiReCT and deep-learning-based segmentation is more sensitive to subvoxel cortical thinning than Freesurfer. While anatomical segmentation of a T1-weighted image now takes seconds, existing implementations of DiReCT rely on iterative image registration methods which can take up to an hour per volume. On the other hand, learning-based deformable image registration methods like VoxelMorph have been shown to be faster than classical methods while improving registration accuracy. This paper proposes CortexMorph, a new method that employs unsupervised deep learning to directly regress the deformation field needed for DiReCT. By combining CortexMorph with a deep-learning-based segmentation model, it is possible to estimate region-wise thickness in seconds from a T1-weighted image, while maintaining the ability to detect cortical atrophy. We validate this claim on the OASIS-3 dataset and the synthetic cortical thickness phantom of Rusak et al.

eess.IV↗

arXiv2020

Simultaneous lesion and neuroanatomy segmentation in Multiple Sclerosis using deep neural networks

Segmentation of white matter lesions and deep grey matter structures is an important task in the quantification of magnetic resonance imaging in multiple sclerosis. In this paper we explore segmentation solutions based on convolutional neural networks (CNNs) for providing fast, reliable segmentations of lesions and grey-matter structures in multi-modal MR imaging, and the performance of these methods when applied to out-of-centre data. We trained two state-of-the-art fully convolutional CNN architectures on the 2016 MSSEG training dataset, which was annotated by seven independent human raters: a reference implementation of a 3D Unet, and a more recently proposed 3D-to-2D architecture (DeepSCAN). We then retrained those methods on a larger dataset from a single centre, with and without labels for other brain structures. We quantified changes in performance owing to dataset shift, and changes in performance by adding the additional brain-structure labels. We also compared performance with freely available reference methods. Both fully-convolutional CNN methods substantially outperform other approaches in the literature when trained and evaluated in cross-validation on the MSSEG dataset, showing agreement with human raters in the range of human inter-rater variability. Both architectures showed drops in performance when trained on single-centre data and tested on the MSSEG dataset. When trained with the addition of weak anatomical labels derived from Freesurfer, the performance of the 3D Unet degraded, while the performance of the DeepSCAN net improved. Overall, the DeepSCAN network predicting both lesion and anatomical labels was the best-performing network examined.

cs.CV↗

arXiv2019

Automatic detection of lesion load change in Multiple Sclerosis using convolutional neural networks with segmentation confidence

The detection of new or enlarged white-matter lesions in multiple sclerosis is a vital task in the monitoring of patients undergoing disease-modifying treatment for multiple sclerosis. However, the definition of 'new or enlarged' is not fixed, and it is known that lesion-counting is highly subjective, with high degree of inter- and intra-rater variability. Automated methods for lesion quantification hold the potential to make the detection of new and enlarged lesions consistent and repeatable. However, the majority of lesion segmentation algorithms are not evaluated for their ability to separate progressive from stable patients, despite this being a pressing clinical use-case. In this paper we show that change in volumetric measurements of lesion load alone is not a good method for performing this separation, even for highly performing segmentation methods. Instead, we propose a method for identifying lesion changes of high certainty, and establish on a dataset of longitudinal multiple sclerosis cases that this method is able to separate progressive from stable timepoints with a very high level of discrimination (AUC = 0.99), while changes in lesion volume are much less able to perform this separation (AUC = 0.71). Validation of the method on a second external dataset confirms that the method is able to generalize beyond the setting in which it was trained, achieving an accuracy of 83% in separating stable and progressive timepoints. Both lesion volume and count have previously been shown to be strong predictors of disease course across a population. However, we demonstrate that for individual patients, changes in these measures are not an adequate means of establishing no evidence of disease activity. Meanwhile, directly detecting tissue which changes, with high confidence, from non-lesion to lesion is a feasible methodology for identifying radiologically active patients.

cs.CV↗

arXiv2019

Few-shot brain segmentation from weakly labeled data with deep heteroscedastic multi-task networks

In applications of supervised learning applied to medical image segmentation, the need for large amounts of labeled data typically goes unquestioned. In particular, in the case of brain anatomy segmentation, hundreds or thousands of weakly-labeled volumes are often used as training data. In this paper, we first observe that for many brain structures, a small number of training examples, (n=9), weakly labeled using Freesurfer 6.0, plus simple data augmentation, suffice as training data to achieve high performance, achieving an overall mean Dice coefficient of $0.84 \pm 0.12$ compared to Freesurfer over 28 brain structures in T1-weighted images of $\approx 4000$ 9-10 year-olds from the Adolescent Brain Cognitive Development study. We then examine two varieties of heteroscedastic network as a method for improving classification results. An existing proposal by Kendall and Gal, which uses Monte-Carlo inference to learn to predict the variance of each prediction, yields an overall mean Dice of $0.85 \pm 0.14$ and showed statistically significant improvements over 25 brain structures. Meanwhile a novel heteroscedastic network which directly learns the probability that an example has been mislabeled yielded an overall mean Dice of $0.87 \pm 0.11$ and showed statistically significant improvements over all but one of the brain structures considered. The loss function associated to this network can be interpreted as performing a form of learned label smoothing, where labels are only smoothed where they are judged to be uncertain.

cs.LG↗

arXiv2005

A variational approach to the partition function of an interacting many body system

For the calculation of the partition function $\mathcal{Z}$ of small, isolated and interacting many body systems an improvement with respect to previous formulations is presented. By including anharmonicities and employing a variational approach quantum effects can be treated even at very low temperatures. In order to test its accuracy the novel approach is applied to the exactly solvable Lipkin-Meshkov-Glick model (LMGM). For thermodynamic properties and level densities good agreement with the exact calculations is found.

nucl-th↗

arXiv2005

Damped collective motion of many body systems: A variational approach to the quantal decay rate

We address the problem of collective motion across a barrier like encountered in fission. A formula for the quantal decay rate is derived which bases on a recently developed variational approach for functional integrals. This formula can be applied to low temperatures that have not been accessible within the former PSPA type approach. To account for damping of collective motion one particle Green functions are dressed with appropriate self-energies.

nucl-th↗

arXiv2004

Damped collective motion of isolated many body systems within a variational approach to functional integrals

Two improvements with respect to previous formulations are presented for the calculation of the partition function $\mathcal{Z}$ of small, isolated and interacting many body systems. By including anharmonicities and employing a variational approach quantum effects can be treated even at very low temperatures. A method is proposed of how to include collisional damping. Finally, our approach is applied to the calculation of the decay rate of metastable systems.

nucl-th↗

arXiv2004

Nuclear quantum transport for barrier problems

A method is presented which allows one to introduce collective coordinates self-consistently, in distinction to the Caldeira-Leggett model. It is demonstrated how the partition function Z for the total nuclear system can be calculated to deduce information both on its level density as well as on the decay rate of unstable modes. For the evaluation of Z different approximations are discussed. A recently developed variational approach turns out superior to the conventional methods that include quantum effects on the level of local RPA. Dissipation is taken into account by applying energy smearing, simulating in this way the coupling to more complicated states. In principle, such a coupling must depend on temperature. Previous calculations along another microscopic approach show this fact to imply an intriguing variation of the transport coefficients of collective motion with T. The relevance of this feature is demonstrated for the thermal fission rate and for the formation probability of super-heavy elements.

nucl-th↗

arXiv2003

Influence of microscopic transport coefficients on the formation probabilities for super-heavy elements

The formation probability is shown to increase by a few orders of magnitude if microscopic transport coefficients are used rather than those of the common macroscopic pictures. Quantum effects in collective dynamics are taken into account through the fluctuating force, as exhibited in diffusion coefficients for a Gaussian process. In the range of temperatures considered here, they turn out to be of lesser importance.

nucl-th↗

arXiv2002

The partition function of an interacting many body system: beyond the perturbed static path approximation

Based on the path integral representation of the partition function of a many body system with separable two body interaction we propose a systematic extension of the perturbed static path approximation (PSPA) to lower temperatures. Thereby, special attention must be paid to instabilities of the classical mean field solution in functional space that cause divergencies within the conventional PSPA. As a result we develop an approximation applicable from high to very low temperatures. These findings are tested against exact results for the archetypical cases of a particle moving in a one dimensional double well and the exactly solvable Lipkin model. In particular, we obtain a very good approximation to the level density of the Lipkin model even at low thermal excitations. Our results may have potential applications in low temperature nuclear physics and mesoscopic systems, e.g. for gap fluctuations in nanoscale superconducting devices previously studied within a PSPA type of approximation. PACS: 5.30.-d, 24.60.-k, 21.10.Ma, 74.25.Bt

cond-mat.stat-mech↗