Our most important research publications

Abstract: EEG/ERP (electroencephalography, event related potential) laboratories produce experimental data and metadata. Large amounts of data and various data formats lead to incompatible results and difficult communication among laboratories. Authors' research group has contributed to the building of a neuroinformatics infrastructure by developing and integrating data management and analytic tools for EEG/ERP research. In addition, a module integrated within the EEG/ERP Portal allowing to process experimental data using analytic tools is developed. An integration of the EEG/ERP Portal with other systems is ensured using web services.

Abstract: This article is focused on development of a tool supporting physicians with an appropriate treatment decisions at patients with acute ischemic stroke. The automated tools for infarction core area delineation could provide important information about the volume of the infarction core. This article describes such automated method results used on both cerebral and perfusion blood volume computed tomography maps compared with manual infarction core delineations made by two physicians.

Abstract: In our contemporary research we are trying to develop the artificial neural network (ANN) model for the prediction of outcome after the occurrence of stroke. This paper mentions some important facts about stroke as well as the urgent need for Computer Assisted Decision Support (CAMS) systems in the relation to clinical practice. The short review of related studies of ANN in medicine is included. The model input and output parameters were selected and are also described. The basic ANN design for the predictive model is mentioned together with the future directions of our research.

Abstract: Attention of drivers is a key factor of road safety. Inattentive drivers are dangerous to their surroundings and cause a considerable number of accidents. However, decline of attention, especially during long rides, is natural and it is worth to know if it can be investigated from electric activity of the brain. Since there is an evidence of relationship between attention and P3 component which is widely used in design and analysis of experiments by the method of event related potentials (ERPs), the ERP experiment based on auditory stimulation was designed and performed on tested subjects, which drove a car simulator on a monotonous track. The background of the used method, experimental design, data processing, and results of the performed experiments are described in this paper.

Abstract: This paper proposes a gradient domain deformation for wrapping surface models of muscles around bones as they move during a simulation of physiological activities. Each muscle is associated with one or more poly-lines that represent the muscle skeleton to which the surface model of the muscle is bound so that transformation of the skeleton (caused by the movement of bones) produces transformation of the vertices of the mesh subject to Laplacian linear constraints to preserve the local shape of the mesh and non-linear volume constraints to preserve the volume of the mesh. All these constraints form a system of equations that is solved using the iterative Gauss-Newton method with Lagrange multipliers. Our C++ implementation can wrap a muscle of medium size in about a couple of ms up to 400 ms on commodity hardware depending on the type of parallelization, whilst it can keep the change in volume below 0.04%. A preliminary biomechanical assessment of the proposed technique suggests that it can produce realistic results and thanks to its rapid processing speed, it might be an attractive alternative to the methods that are used in clinical practise at present.

Abstract: This paper deals with influence of mental load on drivers? attention. EEG (electroencephalography) and ERP (event related potentials) techniques are used for investigation of the level of drivers? attention. A short overview of experiments dealing with drivers? attention is given and the ERP technique and the P3 component are described. General assumptions related to P3 amplitude and P3 latency are introduced. Then experimental design and the course of the experiment are described. The resulting EEG/ERP data are analyzed and their interpretation is provided.

Abstract: "Since the evoked responses obtained by stimulation are much weaker than the continuous electroencephalographic signal, the correct signal analysis enhances stimulation-driven signal components. The paper proposes methods for event-related potential processing for brain-computer interfaces based on matching pursuit. The suggested method is compared with another simple method which is frequently used for feature extraction. A multi-layer perceptron was used for classification. The results can be used to improve the feature extraction for BCI systems."

Abstract: The most indicated imaging technique before operations is computed tomography. Preoperative deliberation requires not only its own evaluation of pathological findings, but also the overall anatomy, perfusion and volume of various anatomical parts, which must be accurately identified. There is a wide range for computer-assisted diagnosis, which should reduce and refine the assessment carried out by man and help better decide on the appropriate therapeutic procedure (resection, multiphase power, chemotherapy, ablation methods), and, among other things, reduce the number of unnecessary surgical procedures while on the other hand enabling the implementation of extensive resection. We aims to create a comprehensive system for computed imaging, which would allow the detection and analysis of focal lesions, determination of their segmental localisation, assessment of perfusion conditions and performance of a virtual resection including the establishment of volume resected and organ remaining. Such a system is not yet commercially available.

Abstract: The MetaMed is a tool for data and meta data extraction primarily from a heterogeneous medical data. It is a non-interactive command line open source application for processing a large amount of data. An extracted meta data is stored in the Resource Description Framework and is structured by OWL ontologies. It enables a semantic interoperability. Meta data as an index data prevents us to solve different file format and version of raw files again and again for a following research. It enables a better performance. Indexed data can be directly used in a user interface and querying for an appropriate raw data can be done. The MetaMed and the research information system is confirmed with research primarily focused on a cerebrovascular diseases.

Abstract: This paper describes a method that represents a muscle by a realistic chaff of muscle fibres that are automatically generated in the volume defined by the surface mesh of the muscle which itself automatically wraps around bones as they move. Our C++ implementation can decompose the volume into muscle fibres, which is done by a slice-by-slice morphing of predefined fibres template into the muscle volume, and visualise the result in only about 1000 ms on commodity hardware. Hence, the method is fast enough to be suitable for interactive educational medical software. Although a biomechanical assessment has yet to be done, we believe that the method could be used also in clinical biomechanical applications to extract information on the current muscle lever arm and fibre path and, thanks to its rapid processing speed, it might be an attractive alternative to current methods.

Abstract: The EEG/ERP (Electroencephalography/Even- Related Potentials) Portal serves for storage and management of ERP experiments. The protocols describing these experiments are designed and implemented using various software applications and/or hardware devices. Some protocols are kept in XML files and stored in the EEG/ERP portal database as LOB (Large Object) data types. Since it is important for researchers to directly find details about design of the ERP protocols in the related XML files and it is also hard to use LOB data for generating the semantic web output, storage types provided by the Oracle 11g database system were investigated and compared. Therefore a new model of storage of ERP protocols is introduced. This model is finally simplified after analysis of results of performance tests.

Abstract: There is a problem with global search and data sharing. Our research group developed the EEG/ERP (electroencephalography, event-related potentials) portal; a system for storage and management of EEG/ERP resources - data, metadata, tools and materials related to EEG/ERP experiments. The EEG/ERP portal and existing solutions of data integration are presented. Authors registered the EEG/ERP portal as a source of neuroscience data and metadata within the world known project - Neuroscience Information Framework (NIF). Developed web services for harvesting EEG/ERP scenarios and experiments within NIF are presented.

Abstract: Glucose is an important source of energy for cells. In clinical practice, we measure glucose level in blood and interstitial fluid. Each method has its pros and cons, and both levels correlate with each other. As the body tries to maintain the glucose level within a particular range to avoid adverse effects, it is desirable to predict future glucose levels in order to aid provided health care. We can see this desire in research, e.g. research on glukose transporters of cells. As yet another example, we can see it with diabetic patients, patients in a metabolic intensive care unit, particularly. In this paper, a glucose level prediction method is proposed.

Abstract: The Kohonen Self-organizing Feature Map (SOM) has been developed for clustering input vectors and for projection of continuous high-dimensional signal to discrete low-dimensional space. The application area, where the map can be also used, is the processing of text documents. Within the project WEBSOM, some methods based on SOM have been developed. These methods are suitable either for text documents information retrieval or for organization of large document collections. All methods have been tested on collections of English and Finnish written documents. This article deals with the application of WEBSOM methods to Czech written documents collections. The basic principles of WEBSOM methods, transformation of text information into the real components feature vector and results of documents classification are described. The Carpenter-Grossberg ART-2 neural network, usually used for adaptive vector clustering, was also tested as a document categorization tool. The results achieved by using this network are also presented.

Abstract: Study deals with attention deficit and hyperactivity disorder with connection to coordination disorder. Aim of study is to assess differences in attention according to motor development in preschool children. Sixteen children with mean age of 5,87 (SD ? 0,55) participated in study. To assess a motor development TGMD - 2 was used. The attention was measured through the electric activity of brain and analysis of evoked potentials. The results confirm assumption that children with coordination disorder to suffer from disorder in visuospatial attention and also attention deficit to acoustic stimuli. Together with worse processing and information storage it can cause of coordination disorder and problems in motor learning as well.

Abstract: Because the Semantic Web uses its technologies for presenting data/metadata on the web and common systems are based on object-oriented languages a need for suitable mapping is emerging. This paper describes the difficulties during transformation of data layer represented by object-oriented code into the semantic web structures (OWL, RDF). Since there is difference between semantic expressivity of these data representations it is necessary to fill this semantic gap. Authors investigate these differences in semantics and provide a preliminary idea to add missing semantics into the Java code using Java annotations. These annotations are consequently processed by the proposed framework. The transformation is demonstrated within the EEG/ERP Portal.

Abstract: The paper introduces a system for EEG/ERP (electroencephalography, event-related potentials) data and metadata storage and processing. Since researchers have difficulties with a suitable long-term storage and management of electrophysiology data the presented system helps them to increase both efficiency and effectiveness of their work by providing the means for the storage, management, search and sharing of EEG/ERP data. The requirements specification including the system context, system requirements, project scope, basic features, system users, and data formats and metadata structures is presented. The database structure is proposed; upload, download and interchange of EEG/ERP data and metadata using the web interface are described. The system architecture, used technologies and final realization are described. Data and metadata search and user accounts including system security management are presented. Additional tools and structures as converters of data formats and semantic web ontology are mentioned.

Abstract: The event-related potentials (ERPs) obtained by stimulation are much weaker than the continuous electroencephalographic(EEG) signal. Therefore, the correct signal analysis is vital to detect the stimulation-driven signal components. This paper proposes the combination of matching pursuit for feature extraction and ART 2 neural network for clustering. Then, clusters are filtered and interpreted according to their statistical properties as ERP components or noise. The suggested method can be used to filter the EEG/ERP signal. Furthermore, its results lead to a method that improves averaging when compared to traditional approaches.

Abstract: The musculoskeletal modelling and simulation is an essential step in the process of looking for an optimal strategy to provide patients su?ering from various musculoskeletal disorders, such as osteoporosis, with better health care. In our previous work, we proposed a deformation method suitable for clinical practise that deforms each muscle represented by a surface mesh according to assigned skeleton action lines and preserves its inner volume at the same time. It is built on combination of linear constraints for surface description together with relation of surface to control skeleton and non-linear constraint of volume preservation. It uses Gauss-Newton based iterative solver to ?nd energy minimum ful?lling these conditions. It gains extra performance from exploiting the coarse outer hull for potentially slow and numerically unstable calculations. It achieves excellent ratios of volume preservation and maintains reasonable times in hundreds of milliseconds for our typical meshes, but since each mesh is deformed independently to others, it is unable to provide deformation of multiple interacting meshes without danger of their mutual intersections. This is why a new modi?cation of the method is introduced in this paper that alters both constraint formulation and the iterative solver algorithm to ?x and prevent major intersections between mesh surfaces. It detects and prevents initial intersection in the starting pose of meshes and then prevents new intersections during the solution by constraining the modi?cation step of meshes. It however still considers importance of volume preservation and tries to minimise e?ect of changes on its maintenance. The method was implemented using C++ language and VTK framework and integrated to our human body framework. The results of application to medical data we use show that despite a few open issues, the proposed technique has its merit.

Abstract: When we are processing large quantity of a data, and/or we are computing complex tasks, computational performance of one computer is not enough. Solution is parallel processing. However proper approach to parallel programming doesn't need to well-known to medical experts or computational tool doesn't have native support for parallel programming. Our goal is to design unified interface, which allows parallel approach to our medical researchers. It must provide support for existing medical experiments and it must provide full interoperability.

Abstract: This paper deals with a suitable method for decomposition of EEG/ERP signal to waveforms which are grouped is such way that one or few groups contain ERP P3 waveforms. At the beginning, the EEG/ERP domain is briefly introduced and essential information about EEG and ERP signals is given. Then, the method for waveforms grouping based on matching pursuit algorithm with Gabor dictionary as a preprocessing method for feature extraction for ART2 neural network is explained in detail. Emphasis is placed on selection of suitable feature extraction method. Comparison of tested feature extraction methods and summarization is given at the end.

Abstract: We provide an overview of the association between vitamin D and some neurological diseases where the correlation has repeatedly been described. The majority of literature refers to cerebrovascular diseases, followed by multiple sclerosis and cognitive disorders. Vitamin D hypovitaminosis might be associated with the diseases directly or it might contribute to the disease risk factors (typically in cerebrovascular events). Vitamin D hypovitaminosis may also play a role in patients with residual functional involvement due to a neurological disorder (movement disorders, lack of self-sufficiency) and worsen functional status owing to muscle weakness, instability and falls.

Abstract: This software solves difficulties related to running of signal processing methods. Although several systems that implement signal processing methods exist, their sharing and remote calling is not satisfactorily solved. This software is a custom server-side approach that provides a powerful plug-in engine for integration of signal processing methods. The plug-in engine ensures high modularity and flexibility of the system. Since the implemented methods are accessible via the SOAP Web Service, integration with another system is available. There is also possible to use the system locally via a web browser. The set of basic methods is already implemented.

Abstract: The EEG/ERP (Electroencephalography/Even- Related Potentials) Portal serves for storage and management of ERP experiments. The protocols describing these experiments are designed and implemented using various software applications and/or hardware devices. Some protocols are kept in XML files and stored in the EEG/ERP portal database as LOB (Large Object) data types. Since it is important for researchers to directly find details about design of the ERP protocols in the related XML files and it is also hard to use LOB data for generating the semantic web output, storage types provided by the Oracle 11g database system were investigated and compared. Therefore a new model of storage of ERP protocols is introduced. This model is finally simplified after analysis of results of performance tests.