Alexander Horsch

• Dr. rer. nat. Dr. med. habil.
• Professor (Full) at UiT The Arctic University of Norway

The field of varying feature space in online learning settings, also known as haphazard inputs, is very prominent nowadays due to its applicability in various fields. However, the current solutions to haphazard inputs are model-dependent and cannot benefit from the existing advanced deep-learning methods, which necessitate inputs of fixed dimension...

Accelerometers are frequently used to assess physical activity in large epidemiological studies. They can monitor movement patterns and cycles over several days under free-living conditions and are usually either worn on the wrist or the hip. While wrist-worn accelerometers have been frequently used to additionally assess sleep and time in bed beha...

Untrimmed videos have interrelated events, dependencies, context, overlapping events, object-object interactions, domain specificity, and other semantics that are worth highlighting while describing a video in natural language. Owing to such a vast diversity, a single sentence can only correctly describe a portion of the video. Dense Video Captioni...

We study the online learning problem characterized by the varying input feature space of streaming data. Although LSTMs have been employed to effectively capture the temporal nature of streaming data, they cannot handle the dimension-varying streams in an online learning setting. Therefore, we propose a dynamic LSTM-based novel method, called packe...

Outer context: The innovation of microscopic and nanoscopic imaging technology has opened several research directions in life science. However, these technologies require larger storage space and challenge automatic analysis. Research question: There is an utmost requirement for storing large-sized images with a smaller storage footprint and corr...

The invention of microscopy- and nanoscopy-based imaging technology opened up different research directions in life science. However, these technologies create the need for larger storage space, which has negative impacts on the environment. This scenario creates the need for storing such images in a memory-efficient way. Compact image representati...

Paradoxical associations have been observed for leisure-time physical activity (LTPA) and occupational physical activity (OPA) and several health-related outcomes. Typically, higher LTPA is associated with health benefits and high OPA with health hazards. Using data from the Tromsø Study (2015-2016), we assessed how questionnaire-based LTPA and OPA...

Structured beams carrying topological defects, namely phase and Stokes singularities, have gained extensive interest in numerous areas of optics. The non-separable spin and orbital angular momentum states of hybridly polarized Stokes singular beams provide additional freedom for manipulating optical fields. However, the characterization of hybridly...

Accelerometers are increasingly used to observe human behavior such as physical activity under free-living conditions. An important prerequisite to obtain reliable results is the correct calibration of the sensors. However, accurate calibration is often neglected, leading to potentially biased results. Here, we demonstrate and quantify the effect o...

Universities are facing many challenges as they are expected to prepare their students in the best possible way to contribute to sustainable societal- and industrial development. Students will become researchers, innovators, entrepreneurs, and role models, and should be able to contribute in the transition to a greener and smarter future. The task...

Non-linear regression modeling is common in epidemiology for prediction purposes or estimating relationships between predictor and response variables. Restricted cubic spline (RCS) regression is one such method, for example, highly relevant to Cox proportional hazard regression model analysis. RCS regression uses third-order polynomials joined at k...

The proliferation of high-content microscopes (\(\sim \)32 GB for a single image) and the increasing amount of image data generated daily have created a pressing need for compact storage solutions. Not only is the storage of such massive image data cumbersome, but it also requires a significant amount of storage and data bandwidth for transmission....

Non-linear regression modeling is common in many fields for prediction purposes or estimating relationships between predictor and response variables. For example, restricted cubic spline regression can model non-linear relationships as third-order polynomials joined at knot points. The standard approach is to place knots by a regular sequence of qu...

Structured beams carrying topological defects, namely phase and Stokes singularities, have gained extensive interest in numerous areas of optics. The non-separable spin and orbital angular momentum states of hybridly polarized Stokes singular beams provide additional freedom for manipulating optical fields. However, the characterization of hybridly...

Data from consumer-based devices for collecting personal health-related data could be useful in diagnostics and treatment. This requires a flexible and scalable software and system architecture to handle the data. This study examines the existing mSpider platform, addresses shortcomings in security and development, and suggests a full risk analysis...

Artificial networks of neurons and algorithms mimic human neural systems intertwined and stacked. Most can’t comprehend a human mind, yet they work flawlessly until they don’t. With millions of billions of parameter optimizations in hundreds of layers over a short time, the question is not ‘if’ something will go wrong, but ‘when’.

When it comes to interpretability, ML models, particularly DL models, are frequently regarded as a black box due to their complexity and lack of transparency in approach. It is fairly simple to train a network to be specific. A DL model learns to classify an object, recognize text, or generate digital images. It efficiently encapsulates feature lea...

Artificial Intelligence (AI) and modern computing captivate a large and growing number of people. It’s fascinating to see how they progressed from a mere impression of mimicking human-like behavior to surpassing human-level performance that fits in one’s pocket.

Previous chapters have discussed the learning, performance and explainability of NNs based on crisp inputs, weights, parameters, training samples and other information pieces. It was demonstrated how a deep network propagates information into its layers.

The chapter commences with a prevalent phrase in the modern era “AI will take over the world!” (Crawford and Calo 2016). There are two major interpretations of the phrase. The first is to acknowledge AI as a developing technology with the ability to cater to larger organizations, global automation, and streamline inefficient procedures. The other i...

Streaming classification methods assume the number of input features is fixed and always received. But in many real-world scenarios, some features are reliable while others are unreliable or inconsistent. We propose a novel online deep learning-based model called Auxiliary Network (Aux-Net), which is scalable and agile and can handle any number of...

Non-linear regression modeling is common in many fields for prediction purposes or estimating relationships between predictor and response variables. For example, restricted cubic spline regression can model non-linear relationships as third-order polynomials joined at knot points. The standard approach is to place knots by a regular sequence of qu...

Non-linear regression modeling is common in many fields for prediction purposes or estimating relationships between predictor and response variables. For example, restricted cubic spline regression can model non-linear relationships as third-order polynomials joined at knot points. The standard approach is to place knots by a regular sequence of qu...

Background Serious public-health concerns such as overweight and obesity are in many cases caused by excess intake of food combined with decreases in physical activity. Smart scales with wireless data transfer can, together with smart watches and trackers, observe changes in the population’s health. They can present us with a picture of our metabol...

Image retrieval has garnered growing interest in recent times. The current approaches are either supervised or self-supervised. These methods do not exploit the benefits of hybrid learning using both supervision and self-supervision. We present a novel Master Assistant Buddy Network (MABNet) for image retrieval which incorporates both learning mech...

Fluorescence microscopy is a quintessential tool for observing cells and understanding the underlying mechanisms of life-sustaining processes of all living organisms. The problem of extracting 3D shape of mitochondria from fluorescence microscopy images remains unsolved due to the complex and varied shapes expressed by mitochondria and the poor res...

Background: Smart scales with wireless data transfer have existed for more than a decade. These smart body composition scales can, together with smart watches and trackers, observe changes in the population's health. Combining body composition data with physical activity measurements from devices such as smart watches could contribute to building a...

Mitochondria play a crucial role in cellular metabolism. This paper presents a novel method to visualize mitochondria in living cells without the use of fluorescent markers. We propose a physics-guided deep learning approach for obtaining virtually labeled micrographs of mitochondria from bright-field images. We integrate a microscope’s point sprea...

Epidemiological literature on the relationship between physical activity and chronic pain is scarce and inconsistent. Hence, our aim was to assess the relationship applying comprehensive methodology, including self-reported and accelerometer measures of physical activity and different severity levels of chronic pain. We used data from the Tromsø St...

Background The Saltin-Grimby Physical Activity Level Scale (SGPALS) is commonly used to measure physical activity (PA) in population studies, but its validity in adolescents is unknown. This study aimed to assess the criterion validity of the SGPALS against accelerometry in a large sample of adolescents. A secondary aim was to examine the validity...

The increased development and use of ubiquitous digital services reinforce the trend where health-related data is generated everywhere. Data usage in different areas introduces different terms for the same or similar concepts. This adds to the confusion of what these terms represent. We aim to provide an overview of concepts and terms used in conne...

We commonly adjust for confounding factors in analytical observational epidemiology to reduce biases that distort the results. Stratification and matching are standard methods for reducing confounder bias. Coarsened exact matching (CEM) is a recent method using stratification to coarsen variables into categorical variables to enable exact matching...

The use of accelerometers has become an established method in population research. Accelerometers are small body-worn sensors that can monitor movement patterns and cycles over several days under free-living conditions. A key requirement for any accelerometer-based analysis is the reliable discrimination of sleep and wake episodes. However, many st...

This study describes a method for extracting the stride parameter ground contact time (GCT) from inertial sensor signals in sprinting. Five elite athletes were equipped with inertial measurement units (IMU) on their ankles and performed 34 maximum 50 and 100-m sprints. The GCT of each step was estimated based on features of the recorded IMU signals...

We often locate ourselves in a trade-off situation between what is predicted and understanding why the predictive modeling made such a prediction. This high-risk medical segmentation task is no different where we try to interpret how well has the model learned from the image features irrespective of its accuracy. We propose image-specific fine-tuni...

To date, non-wear detection algorithms commonly employ a 30, 60, or even 90 mins interval or window in which acceleration values need to be below a threshold value. A major drawback of such intervals is that they need to be long enough to prevent false positives (type I errors), while short enough to prevent false negatives (type II errors), which...

Examining specific sub-cellular structures while minimizing cell perturbation is important in the life sciences. Fluorescence labeling and imaging is widely used for introducing specificity despite its perturbative and photo-toxic nature. With the advancement of deep learning, digital staining routines for label-free analysis have emerged as a repl...

Background The relationship between habitual physical activity (PA) and experimental pain tolerance has been investigated in small samples of young, healthy and/or single‐sex volunteers. We used a large, population‐based sample to assess this relationship in men and women with and without chronic pain. Methods We used data from the sixth and seven...

Objectives: Physical activity may be important in deterring the obesity epidemic. This study aimed to determine whether objectively measured physical activity in first year of upper secondary high school predicted changes in body composition over 2 years of follow-up in a cohort of Norwegian adolescents (n=431). Design: A longitudinal study of a...

Background: Previous studies show large variations in physical activity (PA) levels among adolescents. However, the number of studies is limited and even fewer studies have assessed PA in adolescents by accelerometer devices. This study aimed to describe accelerometer-measured PA levels in adolescents in a population-based cohort in Northern Norwa...

Current non-wear detection algorithms frequently employ a 30- to 90-minute interval in which recorded acceleration needs to be below a threshold value. Such intervals need to be long enough to prevent false positives (type I errors), while short enough to prevent false negatives (type II errors), limiting their ability to achieve a high F1 score. I...

Accurate detection of accelerometer non-wear time is crucial for calculating physical activity summary statistics. In this study, we evaluated three epoch-based non-wear algorithms (Hecht, Troiano, and Choi) and one raw-based algorithm (Hees). In addition, we performed a sensitivity analysis to provide insight into the relationship between the algo...

Objectives We compared the ability of physical activity and sitting time questionnaires (PAQ) for ranking individuals versus continuous volume calculations (physical activity level (PAL), metabolic equivalents of task (MET), sitting hours) against accelerometry measured physical activity as our criterion. Methods Participants in a cohort from the...

Introduction Surveillance of physical activity at the population level increases the knowledge on levels and trends of physical activity, which may support public health initiatives to promote physical activity. Physical activity assessed by accelerometry is challenged by varying data processing procedures, which influences the outcome. We aimed to...

Background: Accelerometers, often in conjunction with heart rate sensors, are extensively used to track physical activity (PA) in research. Research-grade instruments are often expensive and have limited battery capacity, limited storage, and high participant burden. Consumer-based activity trackers are equipped with similar technology and designed...

Background: Previous studies show large variations in physical activity (PA) levels among adolescents. However, the number of studies is limited and even fewer studies have assessed PA in adolescents by accelerometer devices. This study aimed to describe accelerometer-measured PA levels in adolescents in a population-based cohort in Northern Norway...

Effects of low serum 25OHD on age-related changes in muscle mass and function remain unclear. Our aims were to explore associations of baseline 25OHD levels with prevalent and incident sarcopenia and changes in muscle parameters, and to examine the role of parathyroid hormone (PTH) therein. Cross-sectional (n = 975) and prospective analyses (n = 70...

Background: Previous research shows that physical activity has a protective effect on mental distress in adults, but the relationship is less researched and seems more ambiguous for adolescents. Studies in this field have typically been cross-sectional by design and based on self-reported physical activity measures, which are known to be vulnerabl...

Background: Accelerometers, often in conjunction with heart rate sensors, are extensively used to track physical activity (PA) in research. Research-grade instruments are often expensive and have limited battery capacity, limited storage, and high participant burden. Consumer-based activity trackers are equipped with similar technology and designed...

BACKGROUND Accelerometers and combined sensing are extensively used to track physical activity (PA). Research grade instruments are often expensive, have limited battery capacity, limited storage, and high participant burden. Consumer-based activity trackers are equipped with similar technology and are designed for long-time wearing, and can theref...

Accelerometers objectively monitor physical activity, and ongoing research suggests they can also detect patterns of body movement. However, different types of signal (uniaxial, captured by older studies, vs. the newer triaxial) and or/device (validated Actigraph used by older studies, vs. others) may lead to incomparability of results from differe...

Objective: To examine the association between grip strength and history of falls among older individuals, and to assess the possible mediating effect of balance problems on this relationship.Method:Data originate from KORA (Cooperative Health Research in the Region of Augsburg)-Age Study of 808 individuals (65 years and above). Follow-up assessment...

Significant efforts have been devoted to choosing the best configuration of a computing system to run an application energy efficiently. However, available tuning approaches mainly focus on homogeneous systems and are inextensible for heterogeneous systems which include several components (e.g., CPUs, GPUs) with different architectures. This study...

Background: Falls are common among elderly people, and the risk increase with age. Falls are associated with both health and social consequences for the patient, and major societal costs. Identification of risk factors should be investigated to prevent falls. Previous studies have shown anemia to be associated with increased risk of falling, but t...

Background Widespread access to the Internet and an increasing number of Internet users offers the opportunity of using Web-based recalls to collect detailed physical activity data in epidemiologic studies. Objective The aim of this investigation was to evaluate the validity and reliability of a computer-based 24-hour physical activity recall (cpa...

Background and Aims: During infections, many people with diabetes experiences high blood glucose (BG) levels. We propose to utilize patient-gathered information and develop an Electronic Disease Surveillance Monitoring Network, which can possibly detect infectious disease outbreaks during incubation period. The system incorporates data gathered fro...

Background Reduced muscular strength in the old age is strongly related to activity impairment and mortality. However, studies evaluating the gender-specific association between muscularity and mortality among older adults are lacking. Thus, the objective of the present study was to examine gender differences in the association between muscular str...

Objectives: Raw accelerometry is increasingly being used in physical activity research, but diversity in sensor design, attachment and signal processing challenges the comparability of research results. Therefore, efforts are needed to harmonize the methodology. In this article we reflect on how increased methodological harmonization may be achiev...

Skin cancer is one of the most frequently en-countered types of cancer in the Western world. According to the Skin Cancer Foundation Statistics, one in every five Americans develops skin cancer during his/her lifetime. Today, the incurability of advanced cutaneous melanoma raises the importance of its early detection. Since the differentiation of e...

In this epidemiological study we described the characteristics of spatio-temporal gait parameters among a representative, population-based sample of 890 community-dwelling people aged 65 to 90 years. In addition, we investigated the associations between certain gait parameters and a history of falls in study participants. In descriptive analyses sp...

Introduction Accelerometry is an important method for extending our knowledge about intensity, duration, frequency and patterns of physical activity needed to promote health. This study has used accelerometry to detect associations between intensity levels and related activity patterns with multimorbidity and disability. Moreover, the proportion of...

Objectives: The aim of this paper is to discuss how recent developments in the field of big data may potentially impact the future use of wearable sensor systems in healthcare. Methods: The article draws on the scientific literature to support the opinions presented by the IMIA Wearable Sensors in Healthcare Working Group. Results: The followi...

Falls and fractures are among the principal causes of disability, and mortality of older people. Therefore, identifying treatable risk factors for falls in this population is very important.Here we evaluate the association between anemia and falls in community-dwelling people aged 65 years and older. In 2009 967 community-dwelling people aged 65 ye...

Elderly people obtain significant health benefits from physical activity (PA), but the role of activity patterns has scarcely been researched. The present study aims to describe the patterns of PA among different intensities of activity in elderly people. We assess how patterns differ between more and less active groups ('rare', 'average', and 'fre...

Background: Chronic obstructive pulmonary disease (COPD) is a progressive disease affecting the airways, which constitutes a major cause of chronic morbidity and a significant economic and social burden throughout the world. Despite the fact that in COPD patients exacerbations are common acute events causing significant and often fatal worsening o...

Patient diaries as apps on mobile phones are becoming increasingly common, and can be a good support tool for patients who need to organize information relevant for their disease. Self-management is important to achieving diabetes treatment goals and can be a tool for lifestyle changes for patients with Type 2 diabetes. The autoimmune disease Type...

Periodic limb movements are short movements of the legs that can lead to low sleep quality in the general population. Currently, the gold standard to measure periodic limb movements for diagnostic purposes is polysomnography. This is an expensive technique that requires specially fitted laboratories and specialized personnel. In this paper we explo...

13 volunteers walked on a treadmill at 6 different speeds while wearing different accelerometers on different body positions, a spirometry unit and heart rate devices. The goal of this experiment was to find which position on a person's body and which of the sensors is best to measure physical activity (PA) in people by means of an accelerometer.

We measured lower limb activity by means two, above each ankle vertically aligned GT3X+ accelerometers in 4 healthy adults during a single night. Triaxial raw data vector magnitude units (VMU) were recorded at 100 Hz, in order to compare two different modes of operation (i.e. low power mode (LPM)on vs. off).

Due to the large variety of methods for collecting and analyzing accelerometer data to assess physical activity, it is difficult to compare results from different studies. Therefore, this work aims to provide recommendations for the collection and processing of such data in older people, focusing on the assessment of wearing time and intensity leve...

Using wireless body sensor network system (WBSN) based on information and communication technologies (ICTs) to monitor the health status of the elderly gradually becomes popular in the scope of telemedicine. eHealth services based on such technologies can offer healthcare to elderly and chronically ill in and outside their homes, enabling an indepe...

Activity diary for seven days. In the head, ID, anthropometric information, participant’s handedness and hand side of wearing the monitor is documented. For recording the course of day a detailed schema is provided for each day covering from the morning on corner stones of the day. Further, options are provided for time and reason of removing monit...

Graphical illustration of wear time according to diary and NHANES algorithm (left) and time of PA spend in each of the four intensity levels (right) demonstrated in four samples. Left: Graphical illustration of the wear time according to the NHANES algorithm [33] and diary information as a function of time of the day and number of recorded day 1 to...

Result letter for participants. The result letter is thanking the adolescent for participating in the study and is introducing the different levels of physical activity in a common sense. Two diagrams provide the relative time the adolescent has spent in the different levels of activity separately for weekdays and for the weekend. The 60 minutes of...

Surveillance of physical activity (PA) is increasingly based on accelerometry. However, data management guidelines are lacking. We propose an approach for combining accelerometry and diary based PA information for assessment of PA in adolescents and provide an example of this approach using data from German adolescents. The 15-year-old participants...

Introduction: Human body acceleration is often used as an indicator of daily physical activity in epidemiological research. Raw acceleration signals contain three basic components: movement, gravity, and noise. Separation of these becomes increasingly difficult during rotational movements. We aimed to evaluate five different methods (metrics) of p...

Additional information on signal processing and replication of robot findings with simulated data. (DOC)