Refine
Year of publication
Document Type
- Master's Thesis (20)
- Bachelor Thesis (5)
- Preprint (2)
- Diploma Thesis (1)
Language
- English (28) (remove)
Has Fulltext
- yes (28)
Is part of the Bibliography
- no (28)
Keywords
- Application (3)
- Neuronales Netz (3)
- Bildgebung (2)
- MRI (2)
- Maschinelles Lernen (2)
- Simulation (2)
- Software (2)
- stroke (2)
- web application (2)
- 3D Slicer (1)
Institute
- Medizinische Informatik (14)
- Informatik (13)
Implementation of an interactive pattern mining framework on electronic health record datasets
(2019)
Large collections of electronic patient records contain a broad range of clinical information highly relevant for data analysis. However, they are maintained primarily for patient administration, and automated methods are required to extract valuable knowledge for predictive, preventive, personalized and participatory medicine. Sequential pattern mining is a fundamental task in data mining which can be used to find statistically relevant, non-trivial temporal dependencies of events such as disease comorbidities. This works objective is to use this mining technique to identify disease associations based on ICD-9-CM codes data of the entire Taiwanese population obtained from Taiwan’s National Health Insurance Research Database.
This thesis reports the development and implementation of the Disease Pattern Miner – a pattern mining framework in a medical domain. The framework was designed as a Web application which can be used to run several state-of-the-art sequence mining algorithms on electronic health records, collect and filter the results to reduce the number of patterns to a meaningful size, and visualize the disease associations as an interactive model in a specific population group. This may be crucial to discover new disease associations and offer novel insights to explain disease pathogenesis. A structured evaluation of the data and models are required before medical data-scientist may use this application as a tool for further research to get a better understanding of disease comorbidities.
A considerable amount of research in the field of modern robotics deals with mobile agents and their autonomous operation in unstructured, dynamic, and unpredictable environments. Designing robust controllers that map sensory input to action in order to avoid obstacles remains a challenging task. Several biological concepts are amenable to autonomous navigation and reactive obstacle avoidance.
We present an overview of most noteworthy, elaborated, and interesting biologically-inspired approaches for solving the obstacle avoidance problem. We categorize these approaches into three groups: nature inspired optimization, reinforcement learning, and biorobotics. We emphasize the advantages and highlight potential drawbacks of each approach. We also identify the benefits of using biological principles in artificial intelligence in various research areas.
Initial results of an ongoing research in the field of reactive mobile autonomy are presented. The aim is to create a reactive obstacle avoidance method for mobile agent operating in dynamic, unstructured, and unpredictable environment. The method is inspired by the stimulus-response behavior of simple animals. An obstacle avoidance controller is developed that uses raw visual information of the environment. It employs reinforcement learning and is therefore capable of self-developing. This should result with obstacle avoidance behavior that is adaptable and therefore generalizes on various operational modalities. The general assumptions of the agent capabilities, the features of the environment as well as the initial result of the simulation are presented. The plans for improvement and suitable performance evaluation are suggested.
Quantitative assessment of Positron Emission Tomography (PET) imaging can be used for diagnosis and staging of tumors and monitoring of response in cancer treatment. In clinical practice, PET analysis is based on normalized indices such as those based on the Standardized Uptake Value (SUV). Although largely evaluated, these indices are considered quite unstable mainly because of the simplicity of their experimental protocol. Development and validation of more sophisticated methods for the purposes of clinical research require a common open platform that can be used both for prototyping and sharing of the analysis methods, and for their evaluation by clinical users. This work was motivated by the lack of such platform for longitudinal quantitative PET analysis. By following a prototype driven software development approach, an open source tool for quantitative analysis of tumor changes based on multi-study PET image data has been implemented. As a platform for this work, 3D Slicer 4, a free open source software application for medical image computing has been chosen. For the analysis and quantification of PET data, the implemented software tool guides the user through a series of workflow steps. In addition to the implementation of a guided workflow, the software was made extensible by integration of interfaces for the enhancement of segmentation and PET quantification algorithms. By offering extensibility, the PET analysis software tool was transformed into a platform suitable for prototyping and development of PET-specific segmentation and quantification methods. The accuracy, efficiency and usability of the platform were evaluated in reproducibility and usability studies. The results achieved in these studies demonstrate that the implemented longitudinal PET analysis software tool fulfills all requirements for the basic quantification of tumors in PET imaging and at the same time provides an efficient and easy to use workflow. Furthermore, it can function as a platform for prototyping of PET-specific segmentation and quantification methods, which in the future can be incorporated in the workflow.
Segmentation of the Cerebrospinal Fluid from MRI Images for the Treatment of Disc Herniations
(2010)
About 80 percent of people are affected at some point in their lives by lower back pain, which is one of the most common neurological diseases and reasons for long-term disability in the United States. The symptoms are primarily caused by overly heavy lifting and/or overstretching of the back, leading to a rupture and an outward bulge of an intervertebral disc, which puts pressure on and pinches the nerve fibers of the spine. The most common form is a lumbar disc herniation between the fourth and fifth lumbar vertebra and between the fifth lumbar vertebra and the sacrum. In recent years the diagnosis of lower back pain has improved, mainly due to enhanced imaging techniques and imaging quality, but the surgical therapy remains hazardous. Reasons for this include low visibility when accessing the lumbar area and the high risk of causing permanent damage when touching the nerve fibers. A new approach for increasing patient safety is the segmentation and visualization of the cerebrospinal fluid in the lower lumbar region of the vertebral column. For this purpose a new fully-automatic and a semi-automatic approach were developed for separating the cerebrospinal fluid from its surroundings on T2-weighted MRI scans of the lumbar vertebra. While the fully-automatic algorithm is realized by a model-based searching method and a volume-based segmentation, the semi-automatic algorithm requires a seed point and performs the segmentation on individual axial planes through a combination of a region-based segmentation algorithm and a thresholding filter. Both algorithms have been applied to four T2-weighted MRI datasets and are compared with a gold-standard segmentation. The segmentation overlap with the gold-standard was 78.7 percent for the fully-automatic algorithm and 93.1 percent for the semi-automatic algorithm. In the pathological region the fully-automatic algorithm obtained a similarity of 56.6 percent, compared to 87.8 percent for the semi-automatic algorithm.
eHMIS is a Ugandan Hospital Information System (HIS), which targets the Sub-Saharan market. In its first version all forms were programmed statically and adaptations were done by code modifications. In 2014 the development of a second version of eHMIS based on Java started.
This work aims at introducing dynamic forms to this new version. While forms that are significantly important to the workflow of the application will remain static, others are replaced by forms that are dynamically designed by the user. By that, the application will become more flexible and local and situational tailoring will be possible without inducing extra costs.
In this thesis the design, implementation and testing of dynamic forms in eHMIS is discussed. The architecture is based on the questionnaire resource of FHIR®. The module enables the user to create questions and group them into sections and questionnaires. For each question the type of answer expected and other constraints can be defined. A user interface covering all functions was designed, so that no programming skills are required. In a first step dynamic forms were integrated in the application's workflow for recording symptoms, though other fields of application are possible. For testing, a usability experiment was conducted in Tororo Hospital in Eastern Uganda, using the thinking aloud method. Results were analysed and evaluated to detect usability problems and gain a general impression of user satisfaction.
Development and validation of a neural network for adaptive gait cycle detection from kinematic data
(2020)
(1) Background: Instrumented gait analysis is a tool for quantification of the different
aspects of the locomotor system. Gait analysis technology has substantially evolved over
the last decade and most modern systems provide real-time capability. The ability to
calculate joint angles with low delays paves the way for new applications such as real-time
movement feedback, like control of functional electrical stimulation in the rehabilitation
of individuals with gait disorders. For any kind of therapeutic application, the timely
determination of different gait phases such as stance or swing is crucial. Gait phases are
usually estimated based on heuristics of joint angles or time points of certain gait events.
Such heuristic approaches often do not work properly in people with gait disorders due to
the greater variability of their pathological gait pattern. To improve the current state-ofthe-
art, this thesis aims to introduce a data-driven approach for real-time determination
of gait phases from kinematic variables based on long short-term memory recurrent neural
networks (LSTM RNNs).
(2) Methods: In this thesis, 56 measurements with gait data of 11 healthy subjects,
13 individuals with incomplete spinal cord injury and 10 stroke survivors with walking
speeds ranging from 0.2 m
s up to 1 m
s were used to train the networks. Each measurement
contained kinematic data from the corresponding subject walking on a treadmill for 90
seconds. Kinematic data was obtained by measuring the positions of reflective markers on
body landmarks (Helen Hayes marker set) with a sample rate of 60Hz. For constructing a
ground truth, gait data was annotated manually by three raters. Two approaches, direct
regression of gait phases and estimation via detection of the gait events Initial Contact
and Final Contact were implemented for evaluation of the performance of LSTM RNNs.
For comparison of performance, the frequently cited coordinate- and velocity-based event
detection approaches of Zeni et al. were used. All aspects of this thesis have been
implemented within MATLAB Version 9.6 using the Deep Learning Toolbox.
(3) Results: The mean time difference between events annotated by the three raters
was −0.07 ± 20.17ms. Correlation coefficients of inter-rater and intra-rater reliability
yielded mainly excellent or perfect results. For detection of gait events, the LSTM RNN
algorithm covered 97.05% of all events within a scope of 50ms. The overall mean time
difference between detected events and ground truth was −11.62 ± 7.01ms. Temporal
differences and deviations were consistently small over different walking speeds and gait
pathologies. Mean time difference to the ground truth was 13.61 ± 17.88ms for the
coordinate-based approach of Zeni et al. and 17.18 ± 15.67ms for the velocity-based
approach. For estimation of gait phases, the gait phase was determined as a percentage.
Mean squared error to the ground truth was 0.95 ± 0.55% for the proposed algorithm
using event detection and 1.50 ± 0.55% for regression. For the approaches of Zeni et al.,
mean squared error was 2.04±1.23% for the coordinate-based approach and 2.24±1.34%
for the velocity-based approach. Regarding mean absolute error to the ground truth, the
proposed algorithm achieved a mean absolute error of 1.95±1.10% using event detection
and one of 7.25 ± 1.45% using regression. Mean absolute error for the coordinate-based
approach of Zeni et al. was 4.08±2.51% and 4.50±2.73% for the velocity-based approach.
(4) Conclusion: The newly introduced LSTM RNN algorithm offers a high recognition
rate of gait events with a small delay. Its performance outperforms several state-of-theart
gait event detection methods while offering the possibility for real-time processing
and high generalization of trained gait patterns. Additionally, the proposed algorithm
is easy to integrate into existing applications and contains parameters that self-adapt
to individuals’ gait behavior to further improve performance. In respect to gait phase
estimation, the performance of the proposed algorithm using event detection is in line
with current wearable state-of-the-art methods. Compared with conventional methods,
performance of direct regression of gait phases is only moderate. Given the results,
LSTM RNNs demonstrate feasibility regarding event detection and are applicable for
many clinical and research applications. They may be not suitable for the estimation
of gait phases via regression. For LSTM RNNs, it can be assumed, that with a more
optimal configuration of the networks, a much higher performance is achieved.
Aside from hardware, a major component of a Brain Computer Interface is the software that provides the tools for translating raw acquired brain signals into commands to control an application or a device. There’s a range of software, some proprietary, like MATLAB and some free and open source (FOSS), accessible under the GNU General Public License (GNU GPL). OpenViBE is one such freely accessible software. This thesis carries out a functionality and usability test of the platform, looking at its portability, architecture and communication protocols. To investigate the feasibility of reproducing the P300 xDAWN speller BCI presented by OpenViBE, users focused on a character on a 6x6 alphanumeric grid which contained a sequence of random flashes of the rows and columns. Visual stimulus is presented to a user every time the character they are focusing on is highlighted in a row or column. A TMSi analog-to-digital converter was used together with a 32-channel active electrode cap (actiCAP) to record user’s Electroencephalogram (EEG) which was then used in an offline session to train the spatial filter algorithm, and the classifier to identify the P300 evoked potentials, elicited as a user’s reaction to an external stimulus. In an online session, the users tried to spell with the application using the power of their brain signal. Aspects of evoked potentials (EP), both auditory (AEP) and visual (VEP) are further investigated as a validation of results of the P300 speller.
Cytoscape is an open source platform for complex network analysis and visualisation. The Pathway Interaction Database (PID) is a highly structured, curated collection of information about known biomolecular interactions and key cellular processes assembled into signalling pathways. Despite the obvious potential and advantageous usage of both tool (Cytoscape) and information source (PID), there has been no conclusive effort to merge and synergise them. This project aims to make use of the open source characteristics of Cytoscape and optimally visualise the biomolecular interactions found in the PID. This is made possible by the development of a plugin which imports a user-selected pathway file, converts it into a Cytoscape-readable file, and then visualises it. Finally, the user has options to further optimise the pathway by the use of a filter (Barcode – Affymetrix) that removes nodes from the network which are lowly expressed in the Affymetrix microarray data. The user then obtains visual results in a matter of seconds. Additionally, the process of subgraphing nodes through the shortest path method could be applied to the network. This can further assist the user in identifying the molecular pathways of the nodes of interest, a useful feature in network analysis.
Every year, hundreds of thousands of patients are affected by treatment failure or adverse drug reactions, many of which could be revented by pharmacogenomic testing. To address these deficiencies in care, clinics require
automated clinical decision support through computer based systems, which provide clinicians with patient-specific ecommendations. The primary knowledge needed for clinical pharmacogneomics is currently being
developed through textual and unstructured guidelines.
In this thesis, it is evaluated whether a web service can annotate clinically relevant genetic variants with guideline information using web services and identify areas of challenge. The proposed tool displays a formal representation of pharmacogenomic guideline information through a web service and existing resources. It enables the annotation of variant call format (VCF) files with clinical guideline information from the Pharmacogenomic Knowledge Base (PharmGKB) and Clinical Pharmacogenetics Implementation Consortium (CPIC).
The applicability of the web service to nnotate clinically relevant variants with pharmacogenomics guideline information is evaluated by translating five guidelines to a web service workflow and executing the process to annotate publically available genomes. The workflow finds genetic variants covered in CPIC guidelines and influenced drugs.
The results show that the web service could be used to annotate in real time clinically relevant variants with up-to-date pharmacogenomics guideline information, although several challenges such as translating variants into star allele nomenclature and the absence of a unique haplotype nomenclature
remain before the clinical implementation of this approach and the use on other drugs.