Magnetic resonance imaging (MRI) is one of the pillars of modern medical imaging, providing a non-invasive means to generate 3D images of the body with high soft-tissue contrast. Furthermore, the possibilities afforded by the design of MRI sequences enable the signal to be sensitized to a multitude of physiological tissue properties, resulting in a wide variety of distinct MRI modalities for clinical and research use. 

This thesis presents a number of advanced brain MRI applications, which fulfill, to differing extents, two complementary aims. On the one hand, they explore the benefits of a multimodal approach to MRI, combining structural, functional and diffusion MRI, in a variety of contexts. On the other, they emphasize the use of advanced mathematical and computational tools in the analysis of MRI data, such as deep learning, Bayesian statistics, and graph signal processing. 

Paper I introduces an anatomically-adapted extension to previous work in Bayesian spatial priors for functional MRI data, where anatomical information is introduced from a T1-weighted image to compensate for the low anatomical contrast of functional MRI data. 

It has been observed that the spatial correlation structure of the BOLD signal in brain white matter follows the orientation of the underlying axonal fibers. Paper II argues about the implications of this fact on the ideal shape of spatial filters for the analysis of white matter functional MRI data. By using axonal orientation information extracted from diffusion MRI, and leveraging the possibilities afforded by graph signal processing, a graph-based description of the white matter structure is introduced, which, in turn, enables the definition of spatial filters whose shape is adapted to the underlying axonal structure, and demonstrates the increased detection power resulting from their use. 

One of the main clinical applications of functional MRI is functional localization of the eloquent areas of the brain prior to brain surgery. This practice is widespread for various invasive surgeries, but is less common for stereotactic radiosurgery (SRS), a non-invasive surgical procedure wherein tissue is ablated by concentrating several beams of high-energy radiation. Paper III describes an analysis and processing pipeline for functional MRI data that enables its use for functional localization and delineation of organs-at-risk for Elekta GammaKnife SRS procedures. 

Paper IV presents a deep learning model for super-resolution of diffusion MRI fiber ODFs, which outperforms standard interpolation methods in estimating local axonal fiber orientations in white matter. Finally, Paper V demonstrates that some popular methods for anonymizing facial data in structural MRI volumes can be partially reversed by applying generative deep learning models, highlighting one way in which the enormous power of deep learning models can potentially be put to use for harmful purposes. 

Brain activation mapping using functional magnetic resonance imaging (fMRI) has been extensively studied in brain gray matter (GM), whereas in large disregarded for probing white matter (WM). This unbalanced treatment has been in part due to controversies in relation to the nature of the blood oxygenation level-dependent (BOLD) contrast in WM and its detachability. However, an accumulating body of studies has provided solid evidence of the functional significance of the BOLD signal in WM and has revealed that it exhibits anisotropic spatio-temporal correlations and structure-specific fluctuations concomitant with those of the cortical BOLD signal. In this work, we present an anisotropic spatial filtering scheme for smoothing fMRI data in WM that accounts for known spatial constraints on the BOLD signal in WM. In particular, the spatial correlation structure of the BOLD signal in WM is highly anisotropic and closely linked to local axonal structure in terms of shape and orientation, suggesting that isotropic Gaussian filters conventionally used for smoothing fMRI data are inadequate for denoising the BOLD signal in WM. The fundamental element in the proposed method is a graph-based description of WM that encodes the underlying anisotropy observed across WM, derived from diffusion-weighted MRI data. Based on this representation, and leveraging graph signal processing principles, we design subject-specific spatial filters that adapt to a subject’s unique WM structure at each position in the WM that they are applied at. We use the proposed filters to spatially smooth fMRI data in WM, as an alternative to the conventional practice of using isotropic Gaussian filters. We test the proposed filtering approach on two sets of simulated phantoms, showcasing its greater sensitivity and specificity for the detection of slender anisotropic activations, compared to that achieved with isotropic Gaussian filters. We also present WM activation mapping results on the Human Connectome Project’s 100-unrelated subject dataset, across seven functional tasks, showing that the proposed method enables the detection of streamline-like activations within axonal bundles.

Brain activation mapping using functional MRI (fMRI) based on blood oxygenation level-dependent (BOLD) contrast has been conventionally focused on probing gray matter, the BOLD contrast in white matter having been generally disregarded. Recent results have provided evidence of the functional significance of the white matter BOLD signal, showing at the same time that its correlation structure is highly anisotropic, and related to the diffusion tensor in shape and orientation. This evidence suggests that conventional isotropic Gaussian filters are inadequate for denoising white matter fMRI data, since they are incapable of adapting to the complex anisotropic domain of white matter axonal connections. In this paper we explore a graph-based description of the white matter developed from diffusion MRI data, which is capable of encoding the anisotropy of the domain. Based on this representation we design localized spatial filters that adapt to white matter structure by leveraging graph signal processing principles. The performance of the proposed filtering technique is evaluated on semi-synthetic data, where it shows potential for greater sensitivity and specificity in white matter activation mapping, compared to isotropic filtering.

Existing Bayesian spatial priors for functional magnetic resonance imaging (fMRI) data correspond to stationary isotropic smoothing filters that may oversmooth at anatomical boundaries. We propose two anatomically informed Bayesian spatial models for fMRI data with local smoothing in each voxel based on a tensor field estimated from a T1-weighted anatomical image. We show that our anatomically informed Bayesian spatial models results in posterior probability maps that follow the anatomical structure.

Magnetic resonance imaging (MRI) utilizes the magnetic properties of tissues to generate image-forming signals. MRI has exquisite soft-tissue contrast and since tumors are mainly soft-tissues, it offers improved delineation of the target volume and nearby organs at risk. The proposed Magnetic Resonance-only Radiotherapy (MR-only RT) work flow allows for the use of MRI as the sole imaging modality in the radiotherapy (RT) treatment planning of cancer. There are, however, issues with geometric distortions inherent with MR image acquisition processes. These distortions result from imperfections in the main magnetic field, nonlinear gradients, as well as field disturbances introduced by the imaged object. In this thesis, we quantified the effect of system related and patient-induced susceptibility geometric distortions on dose distributions for prostate as well as head and neck cancers. Methods to mitigate these distortions were also studied.

In Study I, mean worst system related residual distortions of 3.19, 2.52 and 2.08 mm at bandwidths (BW) of 122, 244 and 488 Hz/pixel up to a radial distance of 25 cm from a 3T PET/MR scanner was measured with a large field of view (FoV) phantom. Subsequently, we estimated maximum shifts of 5.8, 2.9 and 1.5 mm due to patient-induced susceptibility distortions. VMAT-optimized treatment plans initially performed on distorted CT (dCT) images and recalculated on real CT datasets resulted in a dose difference of less than 0.5%.

 The magnetic susceptibility differences at tissue-metallic,-air and -bone interfaces result in local B₀ magnetic field inhomogeneities. The distortion shifts caused by these field inhomogeneities can be reduced by shimming.  Study II aimed to investigate the use of shimming to improve the homogeneity of local  B₀ magnetic field which will be beneficial for radiotherapy applications. A shimming simulation based on spherical harmonics modeling was developed. The spinal cord, an organ at risk is surrounded by bone and in close proximity to the lungs may have high susceptibility differences. In this region, mean pixel shifts caused by local B₀ field inhomogeneities were reduced from 3.47±1.22 mm to 1.35±0.44 mm and 0.99±0.30 mm using first and second order shimming respectively. This was for a bandwidth of 122 Hz/pixel and an in-plane voxel size of 1×1 mm².  Also examined in Study II as in Study I was the dosimetric effect of geometric distortions on 21 Head and Neck cancer treatment plans. The dose difference in D₅₀ at the PTV between distorted CT and real CT plans was less than 1.0%.

In conclusion, the effect of MR geometric distortions on dose plans was small. Generally, we found patient-induced susceptibility distortions were larger compared with residual system distortions at all delineated structures except the external contour. This information will be relevant when setting margins for treatment volumes and organs at risk.  

The current practice of characterizing MR geometric distortions utilizing spatial accuracy phantoms alone may not be enough for an MR-only radiotherapy workflow. Therefore, measures to mitigate patient-induced susceptibility effects in clinical practice such as patient-specific correction algorithms are needed to complement existing distortion reduction methods such as high acquisition bandwidth and shimming.

Wasserstein Generative Adversarial Networks (WGANs) can be used to generate realistic samples from complicated image distributions. The Wasserstein metric used in WGANs is based on a notion of distance between individual images, which induces a notion of distance between probability distributions of images. So far the community has considered l(2) as the underlying distance. We generalize the theory of WGAN with gradient penalty to Banach spaces, allowing practitioners to select the features to emphasize in the generator. We further discuss the effect of some particular choices of underlying norms, focusing on Sobolev norms. Finally, we demonstrate a boost in performance for an appropriate choice of norm on CIFAR-10 and CelebA.

We propose using the Wasserstein loss for training in inverse problems. In particular, we consider a learned primal-dual reconstruction scheme for ill-posed inverse problems using the Wasserstein distance as loss function in the learning. This is motivated by miss-alignments in training data, which when using standard mean squared error loss could severely degrade reconstruction quality. We prove that training with the Wasserstein loss gives a reconstruction operator that correctly compensates for miss-alignments in certain cases, whereas training with the mean squared error gives a smeared reconstruction. Moreover, we demonstrate these effects by training a reconstruction algorithm using both mean squared error and optimal transport loss for a problem in computerized tomography.

Hypothesis testing has long been a formal and standardized process. Hypothesis generation, on the other hand, remains largely informal. This thesis assess whether eXplainable AI (XAI) can aid in the standardization of hypothesis generation through its utilization as a hypothesis generating tool for medical research. We produce XAI heat maps for a Convolutional Neural Network (CNN) trained to classify Microsatellite Instability (MSI) in colon and gastric cancer with four different XAI methods: Guided Backpropagation, VarGrad, Grad-CAM and Sobol Attribution. We then compare these heat maps with pathology annotations in order to look for differences to turn into new hypotheses. Our CNN successfully generates non-random XAI heat maps whilst achieving a validation accuracy of 85% and a validation AUC of 93% – as compared to others who achieve a AUC of 87%. Our results conclude that Guided Backpropagation and VarGrad are better at explaining high-level image features whereas Grad-CAM and Sobol Attribution are better at explaining low-level ones. This makes the two groups of XAI methods good complements to each other. Images of Microsatellite Insta- bility (MSI) with high differentiation are more difficult to analyse regardless of which XAI is used, probably due to exhibiting less regularity. Regardless of this drawback, our assessment is that XAI can be used as a useful hypotheses generating tool for research in medicine. Our results indicate that our CNN utilizes the same features as our basic pathology annotations when classifying MSI – with some additional features of basic pathology missing – features which we successfully are able to generate new hypotheses with. 

Patients who are treated in an intensive care unit need continuous monitoring in orderfor clinicians to be prepared to intervene should a secondary event occur. For patientstreated at the neurointensive care unit (NICU) who have suffered a subarachnoid hemorrhage (SAH) this secondary event could be ischemia, resulting in a lack of blood flow.Blood flow can be measured using magnetic resonance imaging (MRI). The process is facilitated with a software called NOVA. Repeated measurements can therefore be performedas a way to monitor the patients, which in this context would be referred to as longitudinalmeasurements. As more data can be collected ways of analyzing and visualizing the datain a comprehensible way is needed. The aim of this thesis was therefore to develop and implement a method for analyzing and visualizing the longitudinal MR measurement data.With this aim in mind two research questions were relevant. The first one was how NOVAflow longitudinal measurements can be visualized to simplify interpretation by cliniciansand the second one was in what ways the longitudinal data can be analyzed. A graphicaluser interface (GUI) was created to present the developed analysis and visualization tool.Development of the tool progressed using feedback from supervisors and neurosurgeons.Visualization and analysis was done through plots of blood velocity and blood flow as themain component as well as a 2D vessel map. The final implementation showed multipleexamples of how the longitudinal data could be both visualized and analyzed. The resultstherefore provided a tool to analyze and visualize NOVA flow longitudinal measurementsin a way which was easily interpreted. Further improvements of the tool is possible andan area of improvement could involve increasing the adaptability of the tool.

In this paper, we propose a framework for gradually improving the quality of an already existing image descriptor. The descriptor used in this paper (Afkham et al., 2013) uses the response of a series of discriminative components for summarizing each image. As we will show, this descriptor has an ideal form in which all categories become linearly separable. While, reaching this form is not feasible, we will argue how by replacing a small fraction of these components, it is possible to obtain a descriptor which is, on average, closer to this ideal form. To do so, we initially identify which components do not contribute to the quality of the descriptor and replace them with more robust components. Here, a joint feature selection method is used to find improved components. As our experiments show, this change directly reflects in the capability of the resulting descriptor in discriminating between different categories.

Numerous applications in diffusion MRI involve computing the orientationally-averaged diffusion-weighted signal. Most approaches implicitly assume, for a given b-value, that the gradient sampling vectors are uniformly distributed on a sphere (or shell), computing the orientationally-averaged signal through simple arithmetic averaging. One challenge with this approach is that not all acquisition schemes have gradient sampling vectors distributed over perfect spheres. To ameliorate this challenge, alternative averaging methods include: weighted signal averaging; spherical harmonic representation of the signal in each shell; and using Mean Apparent Propagator MRI (MAP-MRI) to derive a three-dimensional signal representation and estimate its isotropic part. Here, these different methods are simulated and compared under different signal-to-noise (SNR) realizations. With sufficiently dense sampling points (61 orientations per shell), and isotropically-distributed sampling vectors, all averaging methods give comparable results, (MAP-MRI-based estimates give slightly higher accuracy, albeit with slightly elevated bias as b-value increases). As the SNR and number of data points per shell are reduced, MAP-MRI-based approaches give significantly higher accuracy compared with the other methods. We also apply these approaches to in vivo data where the results are broadly consistent with our simulations. A statistical analysis of the simulated data shows that the orientationally-averaged signals at each b-value are largely Gaussian distributed.

Karolinska University Hospital, Huddinge, Sweden, has long desired to plan hip prostheses with Computed Tomography (CT) scans instead of plain radiographs to save time and patient discomfort. This has not been possible previously as their current software is limited to prosthesis planning on traditional 2D X-ray images. The purpose of this project was therefore to create an application (software) that allows medical professionals to derive a 2D image from CT images that can be used for prosthesis planning.

In order to create the application NumPy and The Visualization Toolkit (VTK) Python code libraries were utilised and tied together with a graphical user interface library called PyQt4. The application includes a graphical interface and methods for optimizing the images for prosthesis planning.

The application was finished and serves its purpose but the quality of the images needs to be evaluated with a larger sample group. 

This thesis presents a prototype beacon-based indoor positioning system using IR-based triangulation together with various inertial sensors mounted onto the receiver. By applying a Kalman filter, the mobile receivers can estimate their position by fusing the data received from the two independent measurement systems. Furthermore, the system is aimed to operate and conduct all calculations using microcontrollers. Multiple IR beacons and an AGV were constructed to determine the systems performance.

Empirical and practical experiments show that the proposed localisation system is capable centimeter accuracy. However, because of hardware limitation the system has lacking update frequency and range. With the limitations in mind, it can be established that the final sensor-fused solution shows great promise but requires an extended component assessment and more advanced localisation estimations method such as an Extended Kalman Filter or particle filter to increase reliability.

BACKGROUND: 

Several microsurgical transcranial approaches (MTAs) and endoscopic transnasal approaches (EEAs) to the anterior cranial fossa (ACF) have been described.

OBJECTIVE: 

To provide a preclinical, quantitative, anatomic, comparative analysis of surgical approaches to the ACF.

METHODS: 

Five alcohol-fixed specimens underwent high-resolution computed tomography. The following approaches were performed on each specimen: EEAs (transcribriform, transtuberculum, and transplanum), anterior MTAs (transfrontal sinus interhemispheric, frontobasal interhemispheric, and subfrontal with unilateral and bilateral frontal craniotomy), and anterolateral MTAs (supraorbital, minipterional, pterional, and frontotemporal orbitozygomatic approach). An optic neuronavigation system and dedicated software (ApproachViewer, part of GTx-Eyes II—UHN) were used to quantify the working volume of each approach and extrapolate the exposure of different ACF regions. Mixed linear models with random intercepts were used for statistical analyses.

RESULTS: 

EEAs offer a large and direct route to the midline region of ACF, whose most anterior structures (ie, crista galli, cribriform plate, and ethmoidal roof) are also well exposed by anterior MTAs, whereas deeper ones (ie, planum sphenoidale and tuberculum sellae) are also well exposed by anterolateral MTAs. The orbital roof region is exposed by both anterolateral and lateral MTAs. The posterolateral region (ie, sphenoid wing and optic canal) is well exposed by anterolateral MTAs.

CONCLUSION: 

Anterior and anterolateral MTAs play a pivotal role in the exposure of most anterior and posterolateral ACF regions, respectively, whereas midline regions are well exposed by EEAs. Furthermore, certain anterolateral approaches may be most useful when involvement of the optic canal and nerves involvement are suspected.

Background: Non-invasive diagnostic imaging of atherosclerotic coronary artery disease (CAD) is frequently carried out with cardiovascular magnetic resonance imaging (CMR) or myocardial perfusion single photon emission computed tomography (MPS). CMR is the gold standard for the evaluation of scar after myocardial infarction and MPS the clinical gold standard for ischemia. Magnetic Resonance Imaging (MRI) is at times difficult for patients and may induce anxiety while patient experience of MPS is largely unknown.

Aims: To evaluate image quality in CMR with respect to the sequences employed, the influence of atrial fibrillation, myocardial perfusion and the impact of patient information. Further, to study patient experience in relation to MRI with the goal of improving the care of these patients.

Method: Four study designs have been used. In paper I, experimental cross-over, paper (II) experimental controlled clinical trial, paper (III) psychometric crosssectional study and paper (IV) prospective intervention study. A total of 475 patients ≥ 18 years with primarily cardiac problems (I-IV) except for those referred for MRI of the spine (III) were included in the four studies.

Result: In patients (n=20) with atrial fibrillation, a single shot steady state free precession (SS-SSFP) sequence showed significantly better image quality than the standard segmented inversion recovery fast gradient echo (IR-FGRE) sequence (I). In first-pass perfusion imaging the gradient echo-echo planar imaging sequence (GREEPI) (n=30) had lower signal-to-noise and contrast–to-noise ratios than the steady state free precession sequence (SSFP) (n=30) but displayed a higher correlation with the MPS results, evaluated both qualitatively and quantitatively (II). The MRIAnxiety Questionnaire (MRI-AQ) was validated on patients, referred for MRI of either the spine (n=193) or the heart (n=54). The final instrument had 15 items divided in two factors regarding Anxiety and Relaxation. The instrument was found to have satisfactory psychometric properties (III). Patients who prior CMR viewed an information video scored significantly (lower) better in the factor Relaxation, than those who received standard information. Patients who underwent MPS scored lower on both factors, Anxiety and Relaxation. The extra video information had no effect on CMR image quality (IV).

Conclusion: Single shot imaging in atrial fibrillation produced images with less artefact than a segmented sequence. In first-pass perfusion imaging, the sequence GRE-EPI was superior to SSFP. A questionnaire depicting anxiety during MRI showed that video information prior to imaging helped patients relax but did not result in an improvement in image quality.

Background: Non-invasive diagnostic imaging of atherosclerotic coronary artery disease (CAD) is frequently carried out with cardiovascular magnetic resonance imaging (CMR) or myocardial perfusion single photon emission computed tomography (MPS). CMR is the gold standard for the evaluation of scar after myocardial infarction and MPS the clinical gold standard for ischemia. Magnetic Resonance Imaging (MRI) is at times difficult for patients and may induce anxiety while patient experience of MPS is largely unknown.

Aims: To evaluate image quality in CMR with respect to the sequences employed, the influence of atrial fibrillation, myocardial perfusion and the impact of patient information. Further, to study patient experience in relation to MRI with the goal of improving the care of these patients.

Method: Four study designs have been used. In paper I, experimental cross-over, paper (II) experimental controlled clinical trial, paper (III) psychometric crosssectional study and paper (IV) prospective intervention study. A total of 475 patients ≥ 18 years with primarily cardiac problems (I-IV) except for those referred for MRI of the spine (III) were included in the four studies.

Result: In patients (n=20) with atrial fibrillation, a single shot steady state free precession (SS-SSFP) sequence showed significantly better image quality than the standard segmented inversion recovery fast gradient echo (IR-FGRE) sequence (I). In first-pass perfusion imaging the gradient echo-echo planar imaging sequence (GREEPI) (n=30) had lower signal-to-noise and contrast–to-noise ratios than the steady state free precession sequence (SSFP) (n=30) but displayed a higher correlation with the MPS results, evaluated both qualitatively and quantitatively (II). The MRIAnxiety Questionnaire (MRI-AQ) was validated on patients, referred for MRI of either the spine (n=193) or the heart (n=54). The final instrument had 15 items divided in two factors regarding Anxiety and Relaxation. The instrument was found to have satisfactory psychometric properties (III). Patients who prior CMR viewed an information video scored significantly (lower) better in the factor Relaxation, than those who received standard information. Patients who underwent MPS scored lower on both factors, Anxiety and Relaxation. The extra video information had no effect on CMR image quality (IV).

Conclusion: Single shot imaging in atrial fibrillation produced images with less artefact than a segmented sequence. In first-pass perfusion imaging, the sequence GRE-EPI was superior to SSFP. A questionnaire depicting anxiety during MRI showed that video information prior to imaging helped patients relax but did not result in an improvement in image quality.

Background To assess myocardial perfusion, steady-state free precession cardiac magnetic resonance (SSFP, CMR) was compared with gradient-echo–echo-planar imaging (GRE-EPI) using myocardial perfusion scintigraphy (MPS) as reference. Methods Cardiac magnetic resonance perfusion was recorded in 30 patients with SSFP and in another 30 patients with GRE-EPI. Timing and extent of inflow delay to the myocardium was visually assessed. Signal-to-noise (SNR) and contrast-to-noise (CNR) ratios were calculated. Myocardial scar was visualized with a phase-sensitive inversion recovery sequence (PSIR). All scar positive segments were considered pathologic. In MPS, stress and rest images were used as in clinical reporting. The CMR contrast wash-in slope was calculated and compared with the stress score from the MPS examination. CMR scar, CMR perfusion and MPS were assessed separately by one expert for each method who was blinded to other aspects of the study. Results Visual assessment of CMR had a sensitivity for the detection of an abnormal MPS at 78% (SSFP) versus 91% (GRE-EPI) and a specificity of 58% (SSFP) versus 84% (GRE-EPI). Kappa statistics for SSFP and MPS was 0·29, for GRE-EPI and MPS 0·72. The ANOVA of CMR perfusion slopes for all segments versus MPS score (four levels based on MPS) had correlation r = 0·64 (SSFP) and r = 0·96 (GRE-EPI). SNR was for normal segments 35·63 ± 11·80 (SSFP) and 17·98 ± 8·31 (GRE-EPI), while CNR was 28·79 ± 10·43 (SSFP) and 13·06 ± 7·61 (GRE-EPI). Conclusion GRE-EPI displayed higher agreement with the MPS results than SSFP despite significantly lower signal intensity, SNR and CNR.

This book is a compilation of reviews about the pathogenesis of Type 1 Diabetes. T1D is a classic autoimmune disease. Genetic factors are clearly determinant but cannot explain the rapid, even overwhelming expanse of this disease. Understanding etiology and pathogenesis of this disease is essential. A number of experts in the field have covered a range of topics for consideration that are applicable to researcher and clinician alike. This book provides apt descriptions of cutting edge technologies and applications in the ever going search for treatments and cure for diabetes. Areas including T cell development, innate immune responses, imaging of pancreata, potential viral initiators, etc. are considered.

Background Computed tomography (CT) is an imaging modality commonly used for studies of internal body structures and very useful for detailed studies of body composition. The aim of this study was to develop and evaluate a fully automatic image registration framework for inter-subject CT slice registration. The aim was also to use the results, in a set of proof-of-concept studies, for voxel-wise statistical body composition analysis (Imiomics) of correlations between imaging and non-imaging data.Methods The current study utilized three single-slice CT images of the liver, abdomen, and thigh from two large cohort studies, SCAPIS and IGT. The image registration method developed and evaluated used both CT images together with image-derived tissue and organ segmentation masks. To evaluate the performance of the registration method, a set of baseline 3-single-slice CT images (from 2780 subjects including 8285 slices) from the SCAPIS and IGT cohorts were registered. Vector magnitude and intensity magnitude error indicating inverse consistency were used for evaluation. Image registration results were further used for voxel-wise analysis of associations between the CT images (as represented by tissue volume from Hounsfield unit and Jacobian determinant) and various explicit measurements of various tissues, fat depots, and organs collected in both cohort studies.Results Our findings demonstrated that the key organs and anatomical structures were registered appropriately. The evaluation parameters of inverse consistency, such as vector magnitude and intensity magnitude error, were on average less than 3 mm and 50 Hounsfield units. The registration followed by Imiomics analysis enabled the examination of associations between various explicit measurements (liver, spleen, abdominal muscle, visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), thigh SAT, intermuscular adipose tissue (IMAT), and thigh muscle) and the voxel-wise image information.Conclusion The developed and evaluated framework allows accurate image registrations of the collected three single-slice CT images and enables detailed voxel-wise studies of associations between body composition and associated diseases and risk factors.

Wrist pathologies, particularly fractures common among children and adolescents, present a critical diagnostic challenge. While X-ray imaging remains a prevalent diagnostic tool, the increasing misinterpretation rates highlight the need for more accurate analysis, especially considering the lack of specialized training among many surgeons and physicians. Recent advancements in deep convolutional neural networks offer promise in automating pathology detection in trauma X-rays. However, distinguishing subtle variations between pediatric wrist pathologies in X-rays remains challenging. Traditional manual annotation, though effective, is laborious, costly, and requires specialized expertise. In this paper, we address the challenge of pediatric wrist pathology recognition with a fine-grained approach, aimed at automatically identifying discriminative regions in X-rays without manual intervention. We refine our fine-grained architecture through ablation analysis and the integration of LION. Leveraging Grad-CAM, an explainable AI technique, we highlight these regions. Despite using limited data, reflective of real-world medical study constraints, our method consistently outperforms state-of-the-art image recognition models on both augmented and original (challenging) test sets. Our proposed refined architecture achieves an increase in accuracy of 1.06% and 1.25% compared to the baseline method, resulting in accuracies of 86% and 84%, respectively. Moreover, our approach demonstrates the highest fracture sensitivity of 97%, highlighting its potential to enhance wrist pathology recognition.

During the last few years, segmentation architectures based on deep learning achieved promising results. On the other hand, attention networks have been invented years back and used in different tasks but rarely used in medical applications. This thesis investigated four main attention mechanisms; Squeeze and Excitation, Dual Attention Network, Pyramid Attention Network, and Attention UNet to be used in medical image segmentation. Also, different hybrid architectures proposed by the author were tested. Methods were tested on a kidney tumor dataset and against UNet architecture as a baseline. One version of Squeeze and Excitation attention outperformed the baseline. Original Dual Attention Network and Pyramid Attention Network showed very poor performance, especially for the tumor class. Attention UNet architecture achieved close results to the baseline but not better. Two more hybrid architectures achieved better results than the baseline. The first is a modified version of Squeeze and Excitation attention. The second is a combination between Dual Attention Networks and UNet architecture. Proposed architectures outperformed the baseline by up to 3% in tumor Dice coefficient. The thesis also shows the difference between 2D architectures and their 3D counterparts. 3D architectures achieved more than 10% higher tumor Dice coefficient than 2D architectures.

Large annotated datasets are required for training deep learning models, but in medical imaging data sharing is often complicated due to ethics, anonymization and data protection legislation. Generative AI models, such as generative adversarial networks (GANs) and diffusion models, can today produce very realistic synthetic images, and can potentially facilitate data sharing. However, in order to share synthetic medical images it must first be demonstrated that they can be used for training different networks with acceptable performance. Here, we therefore comprehensively evaluate four GANs (progressive GAN, StyleGAN 1–3) and a diffusion model for the task of brain tumor segmentation (using two segmentation networks, U-Net and a Swin transformer). Our results show that segmentation networks trained on synthetic images reach Dice scores that are 80%–90% of Dice scores when training with real images, but that memorization of the training images can be a problem for diffusion models if the original dataset is too small. Our conclusion is that sharing synthetic medical images is a viable option to sharing real images, but that further work is required. The trained generative models and the generated synthetic images are shared on AIDA data hub.

Medical image analysis has undergone significant advancements with the integration of machine learning techniques, particularly in the realm of bone anomaly detection. The availability of recent datasets and the lack of benchmarking and explainability components provide numerous opportunities in this domain. This study proposes a benchmarking approach to a recently published FracAtlas dataset utilizing state-of-the-art deep-learning models coupled with explainable artificial intelligence (XAI) having two distinct modules. The first module involves the binary classification of fractures in different body parts and explains the decision-making process of the best-performing model using an XAI technique known as EigenCAM. EigenCAM generates heatmaps on every layer of the YOLOv8m model to explain how the model reached a conclusion and localizes the fracture using a heatmap. To verify the heatmap, we also detected fractures using the YOLOv8m detection model, which achieved a mAP@O.5 of 59.5%, outperforming the baseline results on this dataset. The second module involves a multi-class classification task to categorize images into one of the five anatomical regions. The best-performing model for binary classification is the YOLOv8m model, with an accuracy of 83.1%, whereas the best-performing model for multi-class classification is the YOLOv8s, achieving an accuracy of 96.2%.

One of the most vital parts of medical image analysis is the classification of brain tumors. Because tumors are thought to be origins to cancer, accurate brain tumor classification can save lives. As a result, CNN (Convolutional Neural Network)-based techniques for classifying brain cancers are frequently employed. However, there is a problem: CNNs are exposed to vast amounts of training data in order to produce good performance. This is where transfer learning enters into the picture. We present a 4-class transfer learning approach for categorizing Glioma, Meningioma, and Pituitary tumors and non-tumors in this study. The three most prevalent types of brain tumors are glioma, meningioma, and pituitary tumors. Our presented method, which employs the theory of transfer learning, utilizes a pre-trained InceptionResnetV1 method for classifying brain MRI images by extracting features from them using the softmax classifier method. The proposed approach outperforms all prior techniques with a mean classification accuracy of 93.95%. For the evaluation of our method we use kaggle dataset. Precision, recall, and F-score are one of the key performance metrics employed in this study.

Objective

The purpose of this study was to assess if ultra-low-dose CT is a useful clinical alternative to digital radiographs in the evaluation of acute wrist and ankle fractures.

Materials and methods

An ultra-low-dose protocol was designed on a 256-slice multi-detector CT. Patients from the emergency department were evaluated prospectively. After initial digital radiographs, an ultra-low-dose CT was performed. Two readers independently analyzed the images. Also, the radiation dose, examination time, and time to preliminary report was compared between digital radiographs and CT.

Results

In 207 extremities, digital radiography and ultra-low-dose CT detected 73 and 109 fractures, respectively (p < 0.001). The odds ratio for fracture detection with ultra-low-dose CT vs. digital radiography was 2.0 (95% CI, 1.4–3.0). CT detected additional fracture-related findings in 33 cases (15.9%) and confirmed or ruled out suspected fractures in 19 cases (9.2%). The mean effective dose was comparable between ultra-low-dose CT and digital radiography (0.59 ± 0.33 μSv, 95% CI 0.47–0.59 vs. 0.53 ± 0.43 μSv, 95% CI 0.54–0.64). The mean combined examination time plus time to preliminary report was shorter for ultra-low-dose CT compared to digital radiography (7.6 ± 2.5 min, 95% CI 7.1–8.1 vs. 9.8 ± 4.7 min, 95% CI 8.8–10.7) (p = 0.002). The recommended treatment changed in 34 (16.4%) extremities.

Conclusions

Ultra-low-dose CT is a useful alternative to digital radiography for imaging the peripheral skeleton in the acute setting as it detects significantly more fractures and provides additional clinically important information, at a comparable radiation dose. It also provides faster combined examination and reporting times.

A novel 2D particle tracking method, that uses 1) iteration, 2) fast quadratic sub-pixel estimation (with only 28 multiplications per movement), and 3) a previous kernel, has been evaluated and compared with a full-search block-matching method. The comparison with high-frequency ultrasound data (40 MHz) was conducted in silico and on phantoms, which comprised lateral, diagonal, and ellipsoidal movement patterns with speeds of 0–15 mm/s. The mean tracking error was reduced by 68% in silico and 71% for the phantom measurements. When only sub-pixel estimation was used, the decrease in the tracking error was 61% in silico and 57% for the phantom measurements. As well as decreasing the tracking error, the new method only used 70% of the computational time needed by the full-search block-matching method. With a fast method having good tracking ability for high-frequency ultrasound data, we now have a tool to better investigate tissue movements and its dynamic functionality.

The aim of this study was to evaluate tracking performance when an extra reference block is added to a basic block-matching method, where the two reference blocks originate from two consecutive ultrasound frames. The use of an extra reference block was evaluated for two putative benefits: (i) an increase in tracking performance while maintaining the size of the reference blocks, evaluated using in silico and phantom cine loops; (ii) a reduction in the size of the reference blocks while maintaining the tracking performance, evaluated using in vivo cine loops of the common carotid artery where the longitudinal movement of the wall was estimated. The results indicated that tracking accuracy improved (mean - 48%, p<0.005 [in silico]; mean - 43%, p<0.01 [phantom]), and there was a reduction in size of the reference blocks while maintaining tracking performance (mean - 19%, p<0.01 [in vivo]). This novel method will facilitate further exploration of the longitudinal movement of the arterial wall. (C) 2014 World Federation for Ultrasound in Medicine & Biology.

Artificial Intelligence (AI)-based models can help in diagnosing COVID-19 from lung CT scans and X-ray images; however, these models require large amounts of data for training and validation. Many researchers studied Generative Adversarial Networks (GANs) for producing synthetic lung CT scans and X-Ray images to improve the performance of AI-based models. It is not well explored how good GAN-based methods performed to generate reliable synthetic data. This work analyzes 43 published studies that reported GANs for synthetic data generation. Many of these studies suffered data bias, lack of reproducibility, and lack of feedback from the radiologists or other domain experts. A common issue in these studies is the unavailability of the source code, hindering reproducibility. The included studies reported rescaling of the input images to train the existing GANs architecture without providing clinical insights on how the rescaling was motivated. Finally, even though GAN-based methods have the potential for data augmentation and improving the training of AI-based models, these methods fall short in terms of their use in clinical practice. This paper highlights research hotspots in countering the data scarcity problem, identifies various issues as well as potentials, and provides recommendations to guide future research. These recommendations might be useful to improve acceptability for the GAN-based approaches for data augmentation as GANs for data augmentation are increasingly becoming popular in the AI and medical imaging research community.

Objective: In this work, we evaluated a model for the translation of atherosclerotic disease features onto healthy carotid ultrasound images.

Methods: An un-paired domain-to-domain translation model – the cycle Generative Adversarial Network (cycleGAN) – was trained to translate between carotid ultrasound images of healthy arteries and images of pronounced disease. Translation performance was evaluated using the measurement of wall thickness in original and generated images. In addition, we explored disease translation in different tissue segments (subcutaneous tissue, muscle, lumen, far wall, and deep tissues), using structural similarity index measure (SSIM) maps.

Results: Features of pronounced disease were successfully translated to the healthy images (1.2 (0.33) mm vs 0.43 (0.07) mm, p < 0.001), while overall anatomy was retained as SSIM value was equal to 0.78 (0.02). Exploration of translated features showed that both arterial wall and subcutaneous tissues were modified in the translation, but that the subcutaneous tissue was subject to distortion of the anatomy in some cases. The image quality influenced the disease translation performance.

Conclusion: The results show that the model can learn a mapping between healthy and diseased images while retaining the overall anatomical contents. This is the first study on atherosclerosis disease translation in medical images.

Significance: The concept of translating disease onto existing healthy images may serve purposes such as education, cardiovascular risk communication in health conversations, or personalized modelling in precision medicine.

Skeletal muscles are functionally regulated by populations of so-called motor units (MUs). An MU comprises a bundle of muscle fibers controlled by a neuron from the spinal cord. Current methods to diagnose neuromuscular diseases and monitor rehabilitation, and study sports sciences rely on recording and analyzing the bio-electric activity of the MUs. However, these methods provide information from a limited part of a muscle. Ultrasound imaging provides information from a large part of the muscle. It has recently been shown that ultrafast ultrasound imaging can be used to record and analyze the mechanical response of individual MUs using blind source separation. In this work, we present an alternative method - a deep learning pipeline - to identify active MUs in ultrasound image sequences, including segmentation of their territories and signal estimation of their mechanical responses (twitch train). We train and evaluate the model using simulated data mimicking the complex activation pattern of tens of activated MUs with overlapping territories and partially synchronized activation patterns. Using a slow fusion approach (based on 3D CNNs), we transform the spatiotemporal image sequence data to 2D representations and apply a deep neural network architecture for segmentation. Next, we employ a second deep neural network architecture for signal estimation. The results show that the proposed pipeline can effectively identify individual MUs, estimate their territories, and estimate their twitch train signal at low contraction forces. The framework can retain spatio-temporal consistencies and information of the mechanical response of MU activity even when the ultrasound image sequences are transformed into a 2D representation for compatibility with more traditional computer vision and image processing techniques. The proposed pipeline is potentially useful to identify simultaneously active MUs in whole muscles in ultrasound image sequences of voluntary skeletal muscle contractions at low force levels.

Background: Advances in sports medicine, rehabilitation applications and diagnostics of neuromuscular disorders are based on the analysis of skeletal muscle contractions. Recently, medical imaging techniques have transformed the study of muscle contractions, by allowing identifcation of individual motor units’ activity, within the whole studied muscle. However, appropriate image-based simulation models, which would assist the continued development of these new imaging methods are missing. This is mainly due to a lack of models that describe the complex interaction between tissues within a muscle and its surroundings, e.g., muscle fbres, fascia, vasculature, bone, skin, and subcutaneous fat. Herein, we propose a new approach to overcome this limitation.

Methods: In this work, we propose to use deep learning to model the authentic intramuscular skeletal muscle contraction pattern using domain-to-domain translation between in silico (simulated) and in vivo (experimental) image sequences of skeletal muscle contraction dynamics. For this purpose, the 3D cycle generative adversarial network (cycleGAN) models were evaluated on several hyperparameter settings and modifcations. The results show that there were large diferences between the spatial features of in silico and in vivo data, and that a model could be trained to generate authentic spatio-temporal features similar to those obtained from in vivo experimental data. In addition, we used diference maps between input and output of the trained model generator to study the translated characteristics of in vivo data.

Results: This work provides a model to generate authentic intra-muscular skeletal muscle contraction dynamics that could be used to gain further and much needed physiological and pathological insights and assess and overcome limitations within the newly developed research feld of neuromuscular imaging.

Alzheimer disease is one of the most prevalent dementia types affecting elder population. On-time detection of the Alzheimer disease (AD) is valuable for finding new approaches for the AD treatment. Our primary interest lies in obtaining a reliable, but simple and fast model for automatic AD detection. The approach we introduced in the present contribution to identify AD is based on the application of machine learning (ML) techniques. For the first step, we use histogram to transform brain images to feature vectors, containing the relevant "brain" features, which will later serve as the inputs in the classification step. Next, we use the ML algorithms in the classification task to identify AD. The model presented and elaborated in the present contribution demonstrated satisfactory performances. Experimental results suggested that the Random Forest classifier can discriminate the AD subjects from the control subjects. The presented modeling approach, consisting of the histogram as the feature extractor and Random Forest as the classifier, yielded to the sufficiently high overall accuracy rate of 85.77%.

New microscopy techniques are continuously developed, resulting in more rapid acquisition of large amounts of data. Manual analysis of such data is extremely time-consuming and many features are difficult to quantify without the aid of a computer. But with automated image analysis biologists can extract quantitative measurements and increases throughput significantly, which becomes particularly important in high-throughput screening (HTS). This thesis addresses automation of traditional analysis of cell data as well as automation of both image capture and analysis in zebrafish high-throughput screening. 

It is common in microscopy images to stain the nuclei in the cells, and to label the DNA and proteins in different ways. Padlock-probing and proximity ligation are highly specific detection methods that  produce point-like signals within the cells. Accurate signal detection and segmentation is often a key step in analysis of these types of images. Cells in a sample will always show some degree of variation in DNA and protein expression and to quantify these variations each cell has to be analyzed individually. This thesis presents development and evaluation of single cell analysis on a range of different types of image data. In addition, we present a novel method for signal detection in three dimensions. 

HTS systems often use a combination of microscopy and image analysis to analyze cell-based samples. However, many diseases and biological pathways can be better studied in whole animals, particularly those that involve organ systems and multi-cellular interactions. The zebrafish is a widely-used vertebrate model of human organ function and development. Our collaborators have developed a high-throughput platform for cellular-resolution in vivo chemical and genetic screens on zebrafish larvae. This thesis presents improvements to the system, including accurate positioning of the fish which incorporates methods for detecting regions of interest, making the system fully automatic. Furthermore, the thesis describes a novel high-throughput tomography system for screening live zebrafish in both fluorescence and bright field microscopy. This 3D imaging approach combined with automatic quantification of morphological changes enables previously intractable high-throughput screening of vertebrate model organisms.

Microscopy in combination with image analysis has emerged as one of the most powerful and informativeways to analyze cell-based high-throughput screening (HTS) samples in experiments designed to uncover novel drugs and drug targets. However, many diseases and biological pathways can be better studied in whole animals, particularly diseases and pathways that involve organ systems and multicellular interactions, such as organ development, neuronal degeneration and regeneration, cancer metastasis, infectious disease progression and pathogenesis. The zebrafish is a wide-spread and popular vertebrate model of human organfunction and development, and it is unique in the sense that large-scale in vivo genetic and chemical studies are feasible due in part to its small size, optical transparency,and aquatic habitat. To improve the throughput and complexity of zebrafish screens, a high-throughput platform for cellular-resolution in vivo chemical and genetic screens on zebrafish larvae has been developed at Yanik lab at Research Laboratory of Electronics, MIT, USA. The system loads live zebrafish from reservoirs or multiwell plates, positions and rotates them for high-speed confocal imaging of organs,and dispenses the animals without damage. We present two improvements to the described system, including automation of positioning of the animals and a novel approach for brightfield microscopy tomographic imaging of living animals.

Brain tumors reduce life expectancy due to the lack of a cure. Moreover, their diagnosis involves complex and costly procedures such as magnetic resonance imaging (MRI) and lengthy, careful examination to determine their severity. However, the timely diagnosis of brain tumors in their early stages may save a patient’s life. Therefore, this work utilizes MRI with a machine learning approach to diagnose brain tumor severity (glioma, meningioma, no tumor, and pituitary) in a timely manner. MRI Gaussian and nonlinear scale features are extracted due to their robustness over rotation, scaling, and noise issues, which are common in image processing features such as texture, local binary patterns, histograms of oriented gradient, etc. For the features, each MRI is broken down into multiple small 8 × 8-pixel MR images to capture small details. To counter memory issues, the strongest features based on variance are selected and segmented into 400 Gaussian and 400 nonlinear scale features, and these features are hybridized against each MRI. Finally, classical machine learning classifiers are utilized to check the performance of the proposed hybrid feature vector. An available online brain MRI image dataset is utilized to validate the proposed approach. The results show that the support vector machine-trained model has the highest classification accuracy of 95.33%, with a low computational time. The results are also compared with the recent literature, which shows that the proposed model can be helpful for clinicians/doctors for the early diagnosis of brain tumors.

A personalized cardiovascular lumped parameter model of the left-sided heart and thesystemic circulation has been developed by the cardiovascular medicine research groupat Linköping University. It provides information about hemodynamics, some of whichcould otherwise only have been retrieved by invasive measurements. The framework forpersonalizing the model is made using 4D Flow MRI data, containing volumes describinganatomy and velocities in three directions. Thus far, the inputs to this model have beengenerated manually for each subject. This is a slow and tedious process, unpractical touse clinically, and unfeasible for many subjects.This project aims to develop a tool to calculate the inputs and run the model for mul-tiple subjects in an automatic way. It has its basis in 4D Flow MRI data sets segmentedto identify the locations of left atrium (LA), left ventricle (LV), and aorta, along with thecorresponding structures on the right side.The process of making this tool started by calculation of the inputs. Planes were placedin the relevant positions, at the mitral valve, aortic valve (AV) and in the ascending aortaupstream the brachiocephalic branches, and flow rates were calculated through them. TheAV plane was used to calculate effective orifice area of AV and aortic cross-sectional area,while the LV end systolic and end diastolic volumes were extracted form the segmentation.The tool was evaluated by comparison with manually created inputs and outputs,using 9 healthy volunteers and one patient deemed to have normal left ventricular func-tion. The patient was chosen from a subject group diagnosed with chronic ischemic heartdisease, and/or a history of angina, together with fulfillment of the high risk score ofcardiovascular diseases of the European Society of Cardiology. This data was evaluatedusing coefficient of variation, Bland-Altman plots and sum squared error. The tool wasalso evaluated visually on some subjects with pathologies of interest.This project shows that it is possible to calculate inputs fully automatically fromsegmented 4D Flow MRI and run the cardiovascular avatar in an automatic way, withoutuser interaction. The method developed seems to be in good to moderate agreement withthose obtained manually, and could be the basis for further development of the model.

This paper investigates the feasibility of using the periocular region for expression recognition. Most works have tried to solve this by analyzing the whole face. Periocular is the facial region in the immediate vicinity of the eye. It has the advantage of being available over a wide range of distances and under partial face occlusion, thus making it suitable for unconstrained or uncooperative scenarios. We evaluate five different image descriptors on a dataset of 1,574 images from 118 subjects. The experimental results show an average/overall accuracy of 67.0%/78.0% by fusion of several descriptors. While this accuracy is still behind that attained with full-face methods, it is noteworthy to mention that our initial approach employs only one frame to predict the expression, in contraposition to state of the art, exploiting several order more data comprising spatial-temporal data which is often not available.

Current research in iris recognition is moving towards enabling more relaxed acquisition conditions. This has effects on the quality of acquired images, with low resolution being a predominant issue. Here, we evaluate a super-resolution algorithm used to reconstruct iris images based on Eigen-transformation of local image patches. Each patch is reconstructed separately, allowing better quality of enhanced images by preserving local information. Contrast enhancement is used to improve the reconstruction quality, while matcher fusion has been adopted to improve iris recognition performance. We validate the system using a database of 1,872 near-infrared iris images. The presented approach is superior to bilinear or bicubic interpolation, especially at lower resolutions, and the fusion of the two systems pushes the EER to below 5% for down-sampling factors up to a image size of only 13×13.

Deformable image registration is important for cancer treatments that utilize an MR-Linac (a linear accelerator with integrated magnetic resonance imaging). Traditional methods of two- and three-dimensional image registration require equipment-specific, handcrafted approaches and lack the necessary image alignment speed for these applications. We introduce a non-black-box deep learning convolutional variational network for two- and three-dimensional deformable medical image registration, called DIRV-Net, which can learn by utilizing either supervised or unsupervised approaches. The model uses variational networks, fields of experts,    and image pyramids for iterative prediction of displacement fields. Our model demonstrates its ability to register two- and three-dimensional images deformed with complex, smooth deformations to near maximum theoretical accuracy. It also shows generalizability and robustness against various image structures and features, including previously unseen ones.

Currently, quantum walk is being explored as a potential method for analyzing medical images. Taking inspiration from Grady's random walk algorithm for image processing, we have developed an approach that leverages the quantum mechanical advantages inherent in quantum walk to detect and segment medical images. Furthermore, the segmented images have been evaluated in terms of clinical relevance. Theoretically, quantum walk algorithms have the potential to offer a more efficient method for medical image analysis compared to traditional methods of image segmentation, such as classical random walk, which do not rely on quantum mechanics. Within this field, there is significant potential for development, and it is of utmost importance to continue exploring and refining these methods. However, it should be noted that there is a long way to go before this becomes something that can be applied in a clinical environment.

Assessment of tumor-infiltrating lymphocytes (TILs) is increasingly recognized as an integral part of the prognostic workflow in triple-negative (TNBC) and HER2-positive breast cancer, as well as many other solid tumors. This recognition has come about thanks to standardized visual reporting guidelines, which helped to reduce inter-reader variability. Now, there are ripe opportunities to employ computational methods that extract spatio-morphologic predictive features, enabling computer-aided diagnostics. We detail the benefits of computational TILs assessment, the readiness of TILs scoring for computational assessment, and outline considerations for overcoming key barriers to clinical translation in this arena. Specifically, we discuss: 1. ensuring computational workflows closely capture visual guidelines and standards; 2. challenges and thoughts standards for assessment of algorithms including training, preanalytical, analytical, and clinical validation; 3. perspectives on how to realize the potential of machine learning models and to overcome the perceptual and practical limits of visual scoring.

In Situ Transcriptomics (IST) is a set of image-based transcriptomics approaches that enables localisation of gene expression directly in tissue samples. IST techniques produce multiplexed image series in which fluorescent spots are either present or absent across imaging rounds and colour channels. A spot’spresence and absence form a type of barcoded pattern that labels a particular type of mRNA. Therefore, the expression of agene can be determined by localising the fluorescent spots and decode the barcode that they form. Existing IST algorithms usually do this in two separate steps: spot localisation and barcode decoding. Although these algorithms are efficient, they are limited by strictly separating the localisation and decoding steps. This limitation becomes apparent in regions with low signal-to-noise ratio or high spot densities. We argue that an improved gene expression decoding can be obtained by combining these two steps into a single algorithm. This allows for an efficient decoding that is less sensitive to noise and optical crowding. We present IST Decoding by Deconvolution (ISTDECO), a principled decoding approach combining spectral and spatial deconvolution into a single algorithm. We evaluate ISTDECOon simulated data, as well as on two real IST datasets, and compare with state-of-the-art. ISTDECO achieves state-of-the-art performance despite high spot densities and low signal-to-noise ratios. It is easily implemented and runs efficiently using a GPU.ISTDECO implementation, datasets and demos are available online at: github.com/axanderssonuu/istdeco

Virtually all the magnetic resonance imaging (MRI) signal of a human originates from water and fat molecules. By utilizing the property chemical shift the signal can be separated, creating water- and fat-only images. From these images it is possible to calculate quantitative fat fraction (FF) images, where the value of each voxel is equal to the percentage of its signal originating from fat. In papers I and II methods for water–fat signal separation are presented and evaluated.

The method in paper I utilizes a graph-cut to separate the signal and was designed to perform well even for a low signal-to-noise ratio (SNR). The method was shown to perform as well as previous methods at high SNRs, and better at low SNRs.

The method presented in paper II uses convolutional neural networks to perform the signal separation. The method was shown to perform similarly to a previous method using a graph-cut when provided non-undersampled input data. Furthermore, the method was shown to be able to separate the signal using undersampled data. This may allow for accelerated MRI scans in the future.

Brown adipose tissue (BAT) is a thermogenic organ with the main purpose of expending chemical energy to prevent the body temperature from falling too low. Its energy expending capability makes it a potential target for treating overweight/obesity and metabolic dysfunctions, such as type 2 diabetes. The most well-established way of estimating the metabolic potential of BAT is through measuring glucose uptake using ¹⁸F-fludeoxyglucose (¹⁸F-FDG) positron emission tomography (PET) during cooling. This technique exposes subjects to potentially harmful ionizing radiation, and alternative methods are desired. One alternative method is measuring the BAT FF using MRI.

In paper III the BAT FF in 7-year olds was shown to be negatively associated with blood serum levels of the bone-specific protein osteocalcin and, after correction for adiposity, thigh muscle volume. This may have implications for how BAT interacts with both bone and muscle tissue.

In paper IV the glucose uptake of BAT during cooling of adult humans was measured using ¹⁸F-FDG PET. Additionally, their BAT FF was measured using MRI, and their skin temperature during cooling near a major BAT depot was measured using infrared thermography (IRT). It was found that both the BAT FF and the temperature measured using IRT correlated with the BAT glucose uptake, meaning these measurements could be potential alternatives to ¹⁸F-FDG PET in future studies of BAT.

The multilinear least-squares (MLLS) problem is an extension of the linear leastsquares problem. The difference is that a multilinear operator is used in place of a matrix-vector product. The MLLS is typically a large-scale problem characterized by a large number of local minimizers. It originates, for instance, from the design of filter networks. We present a global search strategy that allows for moving from one local minimizer to a better one. The efficiency of this strategy is illustrated by results of numerical experiments performed for some problems related to the design of filter networks.

The multilinear least-squares (MLLS) problem is an extension of the linear least-squares problem. The difference is that a multilinearoperator is used in place of a matrix-vector product. The MLLS istypically a large-scale problem characterized by a large number of local minimizers. It originates, for instance, from the design of filter networks. We present a global search strategy that allows formoving from one local minimizer to a better one. The efficiencyof this strategy isillustrated by results of numerical experiments performed forsome problems related to the design of filter networks.

Filter networks is a powerful tool used for reducing the image processing time, while maintaining its reasonably high quality.They are composed of sparse sub-filters whose low sparsity ensures fast image processing.The filter network design is related to solvinga sparse optimization problem where a cardinality constraint bounds above the sparsity level.In the case of sequentially connected sub-filters, which is the simplest network structure of those considered in this paper, a cardinality-constrained multilinear least-squares (MLLS) problem is to be solved. If to disregard the cardinality constraint, the MLLS is typically a large-scale problem characterized by a large number of local minimizers. Each of the local minimizers is singular and non-isolated.The cardinality constraint makes the problem even more difficult to solve.An approach for approximately solving the cardinality-constrained MLLS problem is presented.It is then applied to solving a bi-criteria optimization problem in which both thetime and quality of image processing are optimized. The developed approach is extended to designing filter networks of a more general structure. Its efficiency is demonstrated by designing certain 2D and 3D filter networks. It is also compared with the existing approaches.