Publications by Magnus Borga

Objective: Increase of resistin and/or reduction of adiponectin have been implicated in the development of insulin resistance following weight gain. We aimed to study this prospectively in humans.

Design: Prospective and interventional with parallel control group.

Methods: Twelve healthy men and six healthy women (age 26±6.6 years) and an age-matched control group were recruited. Subjects in the intervention group aimed for a bodyweight increase of 5–15% by doubling the baseline caloric intake by eating at least two fast food-based meals a day in combination withadoption of a sedentary lifestyle for 4 weeks.

Results: Bodyweight increased from 67.6±9.1 to 74.0±11 kg, P<0.001, by the intervention. Insulin levels increased (before: 27.4±12 pmol/l, after: 53.0±22 pmol/l, P=0.004), while plasma levels of adiponectin (before: 5038±3736 ng/ml, after: 6739±7949 ng/ml, P=0.18) and resistin (before: 21.8±19 ng/ml, after: 14.4±6.8 ng/ml, P=0.074) remained unchanged by the weight gain and were similar as in controls. On the other hand, leptin levels increased about threefold following the intervention (before: 5.7±7.4, after: 16±20ng/ml, P=0.008), and also the inflammatory marker C-reactive protein (CRP) increased from 0.34±0.44 to 0.71±0.87 mg/l, P=0.03, when two outliers >10 mg/l were disregarded.

Conclusions: Hyperalimentation reduces insulin sensitivity when weight gain of 9% was combined with reduction of exercise. However, the levels of resistin and adiponectin were unaffected by the intervention,while CRP levels increased within this short time period suggesting that low-grade inflammation can occur early in the process of developing a metabolic syndrome.

We prospectively studied the effects of fast food-based hyperalimentation on insulin sensitivity and components of the metabolic syndrome and analyzed this with respect to sex. Twelve nonobese men and 6 nonobese women (26 +/- 6.6 years old), and an age-matched control group were recruited. Subjects in the intervention group aimed for 5% to 15% weight increase by doubling their regular caloric intake based on at least 2 fast food meals a day while also adopting a sedentary lifestyle for 4 weeks (andlt;5000 steps a day). Weight of Subjects in the intervention group increased from 67.6 +/- 9.1 to 74.0 +/- 11 kg (P andlt;.001), with no sex difference with regard to this or with respect to changes of total abdominal fat volumes or waist circumferences. Fasting insulin (men: before, 3.8 +/- 1.7 mu U/mL, after, 7.4 +/- 3.1 mu U/mL; P=.004; women: before, 4.9 +/- 2.3 mu U/mL; after, 5.9 +/- 2.8 mu U/mL; P =.17), systolic blood pressure (men: before, 117 +/- 13 mm Hg; after, 127 +/- 9.1 mm Hg; P =.002; women: before, 102 +/- 5.1 mm Hg; after, 98 +/- 5.4 mm Hg; P =.39), serum low-density lipoprotein cholesterol, and apolipoprotein B increased only in the men of the intervention group. The sex differences in the metabolic responses to the intervention were linked to a considerable difference in the fat accumulation pattern; 41.4% +/- 9.2% of the increase of the fat volume in the abdominal region was accumulated intraabdominally in men and 22.7 +/- 6.5% in women (P andlt;.0001). This Study thus showed that women are protected, compared with men, against developing intraabdominal obesity when adopting a standardized obesity-provoking lifestyle. Our findings suggest that it is not different lifestyles and/or behaviors that underlie the fact that men have a higher cardiovascular risk at the same level of percentage of body fat than women.

The interest for surgery simulator systems with anatomical models generated from authentic patient data is growing as these systems evolve.With access to volumetric patient data, e.g., from a computer tomography scan, haptic and visual feedback can be created directly from this dataset. This opens the door for patient specific simulations. Hip fracture surgery is one area where simulator systems is useful to train new surgeons and plan operations. To simulate the drilling procedure in this type of surgery, a repositioning of the fractured bone into correct position is first needed. This requires a segmentation process in which the bone segments are identified and the position of the dislocated part is determined. The segmentation must be automatic to cope with the large amount of data from the computer tomography scan. This work presents the first steps in the development of a hip fracture surgery simulation with patient specific models. Visual and haptic feedback is generated from the computer tomography data by simulating fluoroscopic images and the drilling process. We also present an automatic segmentation method to identify the fractured bone and determine the dislocation. This segmentation method is based on nonrigid registration with the Morphon method.

Purpose: To assess the ability of a conventional density mask method to detect mild emphysema by high- resolution computed tomography ( HRCT), to analyze factors influencing quantification of mild emphysema, and to validate a new algorithm for detection of mild emphysema. Material and Methods: Fifty- five healthy male smokers and 34 never- smokers, 61 - 62 years of age, were examined. Emphysema was evaluated visually, by the conventional density mask method, and by a new algorithm compensating for the effects of gravity and artifacts due to motion and the reconstruction algorithm. Effects of the reconstruction algorithm, slice thickness, and various threshold levels on the outcome of the density mask area were evaluated. Results: Forty- nine percent of the smokers had mild emphysema. The density mask area was higher the thinner the slice irrespective of the reconstruction algorithm and threshold level. The sharp algorithm resulted in increased density mask area. The new reconstruction algorithm could discriminate between smokers with and those without mild emphysema, whereas the density mask method could not. The diagnostic ability of the new algorithm was dependent on lung level. At about 90% specificity, sensitivity was 65 - 100% in the apical levels, but low in the rest of the lung. Conclusion: The conventional density mask method is inadequate for detecting mild emphysema, while the new algorithm improves the diagnostic ability but is nevertheless still imperfect.

This article addresses the impact that colored noise, temporal filtering, and temporal detrending have on the fMRI analysis situation. Specifically, it is shown why the detection of event-related designs benefit more from pre-whitening than blocked designs in a colored noise structure. Both theoretical and empirical results are provided. Furthermore, a novel exploratory method for producing drift models that efficiently capture trends and drifts in the fMRI data is introduced. A comparison to currently employed detrending approaches is presented. It is shown that the novel exploratory model is able to remove a major part of the slowly varying drifts that are abundant in fMRI data. The value of such a model lies in its ability to remove drift components that otherwise would have contributed to a colored noise structure in the voxel time series. © 2004 Elsevier Inc. All rights reserved.

This article introduces novel and fundamental improvements of fMRI data analysis. Central is a technique termed constrained canonical correlation analysis, which can be viewed as a natural extension and generalization of the popular general linear model method. The concept of spatial basis filters is presented and shown to be a very successful way of adaptively filtering the fMRI data. A general method for designing suitable hemodynamic response models is also proposed and incorporated into the constrained canonical correlation approach. Results that demonstrate how each of these parts significantly improves the detection of brain activity, with a computation time well within limits for practical use, are provided. © 2003 Elsevier Science (USA). All rights reserved.

A novel method for detecting neural activity in functional magnetic resonance imaging (fMRI) data is introduced. It is based on canonical correlation analysis (CCA), which is a multivariate extension of the univariate correlation analysis widely used in fMRI. To detect homogeneous regions of activity, the method combines a subspace modeling of the hemodynamic response and the use of spatial relationships. The spatial correlation that undoubtedly exists in fMR images is completely ignored when univariate methods such as as t-tests, F-tests, and ordinary correlation analysis are used. Such methods are for this reason very sensitive to noise, leading to difficulties in detecting activation and significant contributions of false activations. In addition, the proposed CCA method also makes it possible to detect activated brain regions based not only on thresholding a correlation coefficient, but also on physiological parameters such as temporal shape and delay of the hemodynamic response. Excellent performance on real fMRI data is demonstrated.

Many applications in image analysis include the problem of linear structure detection, e.g. segmentation of blood vessels in medical images, roads in satellite images, etc. A simple and efficient solution is to apply linear filters tuned to the structures of interest and extract line and edge positions from the filter output. However, if the filter is not carefully designed, artifacts such as ringing can distort the results and hinder a robust detection. In this paper, we study the ringing effects using a common Gabor filter for linear structure detection, and suggest a method for generating non-ring filters in 2D and 3D. The benefits of the non-ring design are motivated by results on both synthetic and natural images.

This paper introduces a novel method for noise reduction in medical images based on concepts of the Non-Local Means algorithm. The main objective has been to develop a method that optimizes the processing speed to achieve practical applicability without compromising the quality of the resulting images. A database consisting of prototypes, composed of pixel neighborhoods originating from several images of similar motif, has been created. By using a dedicated data structure, here Locality Sensitive Hashing (LSH), fast access to appropriate prototypes is granted. Experimental results show that the proposed method can be used to provide noise reduction with high quality results in a fraction of the time required by the Non-local Means algorithm.

There is an increasing demand for imaging methods that can be used for automatic, accurate and quantitative determination of the amounts of abdominal fat. Such methods are important as they will allow the evaluation of some of the risk factors underlying the ’metabolic syndrome’. The metabolic syndrome is becoming common in large parts of the world, and it appears that a dominant risk factor for developing this syndrome is abdominal obesity. Subjects that are afflicted with the metabolic syndrome are exposed to a high risk for developing a large range of diseases such as type 2 diabetes, cardiac failure, and stroke. The aim of this work

Segmentation of images is often posed as a variational problem. As such, it is solved by formulating an energy functional depending on a contour and other image derived terms. The solution of the segmentation problem is the contour which extremizes this functional. The standard way of solving this optimization problem is by gradient descent search in the solution space, which typically suffers from many unwanted local optima and poor convergence. Classically, these problems have been circumvented by modifying the energy functional. In contrast, the focus of this paper is on alternative methods for optimization. Inspired by ideas from the machine learning community, we propose segmentation based on gradient descent with momentum. Our results show that typical models hampered by local optima solutions can be further improved by this approach. We illustrate the performance improvements using the level set framework.

Level set methods are a popular way to solve the image segmentation problem in computer image analysis. A contour is implicitly represented by the zero level of a signed distance function, and evolved according to a motion equation in order to minimize a cost function. This function defines the objective of the segmentation problem and also includes regularization constraints. Gradient descent search is the de facto method used to solve this optimization problem. Basic gradient descent methods, however, are sensitive for local optima and often display slow convergence. Traditionally, the cost functions have been modified to avoid these problems. In this work, we instead propose using a modified gradient descent search based on resilient propagation (Rprop), a method commonly used in the machine learning community. Our results show faster convergence and less sensitivity to local optima, compared to traditional gradient descent.

n/a

The segmentation and analysis of blood vessels hasreceived much attention in the research community. Theresults aid numerous applications for diagnosis andtreatment of vascular diseases. Here we use level setpropagation with local phase information to capture theboundaries of vessels. The basic notion is that localphase, extracted using quadrature filters, allows us todistinguish between lines and edges in an image. Notingthat vessels appear either as lines or edge pairs, weintegrate multiple scales and capture information aboutvessels of varying width. The outcome is a “global”phase which can be used to drive a contour robustly towardsthe vessel edges. We show promising results in2D and 3D. Comparison with a related method givessimilar or even better results and at a computationalcost several orders of magnitude less. Even with verysparse initializations, our method captures a large portionof the vessel tree.

This paper presents a novel method for phase unwrapping for phase sensitive reconstruction in MR imaging. The unwrapped phase is obtained by integrating the phase gradient by solving a Poisson equation. An efficient solver, which has been made publicly available, is used to solve the equation. The proposed method is demonstrated on a fat quantification MRI task that is a part of a prospective study of fat accumulation. The method is compared to a phase unwrapping method based on region growing. Results indicate that the proposed method provides more robust unwrapping. Unlike region growing methods, the proposed method is also straight-forward to implement in 3D.

This paper describes a fully automatic approach for detecting emphysema in CT im ages of the lungs. The method combines an image processing step, where potential emphysematous area s are extracted, and a neural network step trained to rec

A computationally efficient data-driven method for exploratory analysis of functional MRI data is presented. The basic idea is to reveal underlying components in the fMRI data that have maximum autocorrelation. The tool for accomplishing this task is Canonical Correlation Analysis. The proposed method is more robust and much more computationally efficient than independent component analysis, which previously has been applied in fMRI.

We present a novel method that finds edges between certain image features, e.g. gray-levels, and disregards edges between other features. The method uses a channel representation of the features and performs normalized convolution using the channel values as certainties. This means that areas with certain features can be disregarded by the edge filter. The method provides an important new tool for finding tissue specific edges in medical images, as demonstrated by an MR-image example.

This paper presents a general strategy for automated generation of efficient representations in vision. The approach is highly task oriented and what constitutes the relevant information is defined by a set of examples. The examples are pairs of situations that are dependent through the chosen feature but are otherwise independent. Particularly important concepts in the work are mutual information and canonical correlation. How visual operators and representations can be generated from examples are presented for a number of features, e.g. local orientation, disparity and motion. Interesting similarities to biological vision functions are observed. The results clearly demonstrates the potential of combining advanced filtering techniques and learning strategies based on canonical correlation analysis (CCA).

This paper presents a general strategy for designing efficient visual operators. The approach is highly task oriented and what constitutes the relevant information is defined by a set of examples. The examples are pairs of images displaying a strong dependence in the chosen feature but are otherwise independent. Particularly important concepts in the work are mutual information and canonical correlation. Visual operators learned from examples are presented, e.g. local shift invariant orientation operators and image content invariant disparity operators. Interesting similarities to biological vision functions are observed.

This paper presents a novel algorithm that uses CCA and phase analysis to detect the disparity in stereo images. The algorithm adapts filters in each local neighbourhood of the image in a way which maximizes the correlation between the filtered images. The adapted filters are then analysed to find the disparity. This is done by a simple phase analysis of the scalar product of the filters. The algorithm can even handle cases where the images have different scales. The algorithm can also handle depth discontinuities and give multiple depth estimates for semitransparent images.

This paper presents a novel algorithm that uses CCA and phase analysis to detect the disparity in stereo images. The algorithm adapts filters in each local neighbourhood of the image in a way which maximizes the correlation between the filtered images. The adapted filters are then analyzed to find the disparity. This is done by a simple phase analysis of the scalar product of the filters. The algorithm can even handle cases where the images have different scales. The algorithm can also handle depth discontinuities and give multiple depth estimates for semi-transparent images.

A hierarchical representation of the input-output transition function in a learning system is suggested. The choice of either representing the knowledge in a learning system as a discrete set of input-output pairs or as a continuous input-output transition function is discussed. The conclusion that both representations could be efficient, but at different levels of abstraction is made. The difference between strategies and actions is defined. An algorithm for using adaptive critic methods in a two-level reinforcement learning system is presented. Simulations of a one dimensional hierarchical reinforcement learning system is presented.

The subject of this dissertation is to show how learning can be used for multidimensional signal processing, in particular computer vision. Learning is a wide concept, but it can generally be defined as a system’s change of behaviour in order to improve its performance in some sense.

Learning systems can be divided into three classes: supervised learning, reinforcement learning and unsupervised learning. Supervised learning requires a set of training data with correct answers and can be seen as a kind of function approximation. A reinforcement learning system does not require a set of answers. It learns by maximizing a scalar feedback signal indicating the system’s performance. Unsupervised learning can be seen as a way of finding a good representation of the input signals according to a given criterion.

In learning and signal processing, the choice of signal representation is a central issue. For high-dimensional signals, dimensionality reduction is often necessary. It is then important not to discard useful information. For this reason, learning methods based on maximizing mutual information are particularly interesting.

A properly chosen data representation allows local linear models to be used in learning systems. Such models have the advantage of having a small number of parameters and can for this reason be estimated by using relatively few samples. An interesting method that can be used to estimate local linear models is canonical correlation analysis (CCA). CCA is strongly related to mutual information. The relation between CCA and three other linear methods is discussed. These methods are principal component analysis (PCA), partial least squares (PLS) and multivariate linear regression (MLR). An iterative method for CCA, PCA, PLS and MLR, in particular low-rank versions of these methods, is presented.

A novel method for learning filters for multidimensional signal processing using CCA is presented. By showing the system signals in pairs, the filters can be adapted to detect certain features and to be invariant to others. A new method for local orientation estimation has been developed using this principle. This method is significantly less sensitive to noise than previously used methods.

Finally, a novel stereo algorithm is presented. This algorithm uses CCA and phase analysis to detect the disparity in stereo images. The algorithm adapts filters in each local neighbourhood of the image in a way which maximizes the correlation between the filtered images. The adapted filters are then analysed to find the disparity. This is done by a simple phase analysis of the scalar product of the filters. The algorithm can even handle cases where the images have different scales. The algorithm can also handle depth discontinuities and give multiple depth estimates for semi-transparent images.

This paper presents our general strategy for designing learning machines as well as a number of particular designs. The search for methods allowing a sufficient level of adaptivity are based on two main principles: 1. Simple adaptive local models and 2. Adaptive model distribution. Particularly important concepts in our work is mutual information and canonical correlation. Examples are given on learning feature descriptors, modeling disparity, synthesis of a global 3-mode model and a setup for reinforcement learning of online video coder parameter control.

This paper presents a novel algorithm for finding the solution of the generalized eigenproblem where the matrices involved contain expectation values from stochastic processes. The algorithm is iterative and sequential to its structure and uses on-line stochastic approximation to reach an equilibrium point. A quotient between two quadratic forms is suggested as an energy function for this problem and is shown to have zero gradient only at the points solving the eigenproblem. Furthermore it is shown that the algorithm for the generalized eigenproblem can be used to solve three important problems as special cases. For a stochastic process the algorithm can be used to find the directions for maximal variance, covariance, and canonical correlation as well as their magnitudes.

This paper presents a novel learning algorithm that finds the linear combination of one set of multi-dimensional variates that is the best predictor, and at the same time finds the linear combination of another set which is the most predictable. This relation is known as the canonical correlation and has the property of being invariant with respect to affine transformations of the two sets of variates. The algorithm successively finds all the canonical correlations beginning with the largest one. It is shown that canonical correlations can be used in computer vision to find feature detectors by giving examples of the desired features. When used on the pixel level, the method finds quadrature filters and when used on a higher level, the method finds combinations of filter output that are less sensitive to noise compared to vector averaging.

A robust, general and computationally simple reinforcement learning system is presented. It uses a channel representation which is robust and continuous. The accumulated knowledge is represented as a reward prediction function in the outer product space of the input- and output channel vectors. Each computational unit generates an output simply by a vector-matrix multiplication and the response can therefore be calculated fast. The response and a prediction of the reward are calculated simultaneously by the same system, which makes TD-methods easy to implement if needed. Several units can cooperate to solve more complicated problems. A dynamic tree structure of linear units is grown in order to divide the knowledge space into a sufficiently number of regions in which the reward function can be properly described. The tree continuously tests split- and prune criteria in order to adapt its size to the complexity of the problem.