|
Lääketieteellisen fysiikan ja tekniikan yhdistys (LFTY) Finnish Society for Medical Physics and Medical Engineering |
LFT-päivä 12.2.2009, Kuopion yliopisto
Posterikilpailuun osallistuneet
1. Ahokas Sari, Tampereen teknillinen yliopistoDevelopment of low noise active electrodes for EEG measurements.
2. Finnilä Mikko, Oulun yliopisto
Effects of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin Exposure on Bone Material Properties from Macrostructure to Nanostructure.
3. Henelius Andreas, Teknillinen korkeakoulu
Cardiovascular Metrics of Autonomic Nervous System Response to Mental Effort.
4. Holli Kirsi, Tampereen teknillinen yliopisto
Detection of characteristic texture parameters in breast MRI.
5. Koho Sami, Turun yliopisto
Design of a micro volume fluorescence measurement device for immunoassays, using up-converting phosphor labels.
6. Lahtinen Ella, Tampereen teknillinen yliopisto
An Assessment Tool for Assistive Technology.
7. Nieminen Jaakko, Teknillinen korkeakoulu
Enhancement of MRI with SQUID arrays and by polarization encoding.
8. Vuorela Timo, Tampereen teknillinen yliopisto
Unraveling lipoprotein structure via molecular dynamics.
Postereiden abstraktit
Development of low noise active electrodes for EEG measurements
Ahokas Sari, Tampereen teknillinen yliopisto
Nowadays many brain investigation methods exist. EEG measures the electrical activity of the brain and magnetoencephalography (MEG) detects the magnetic fields that are a result of the electrical activity. Computed tomography (CT) describes the brain anatomy. Functional Magnetic Resonance Imaging (fMRI) detects the changes in the blood flow after nerve cell activation. Spectral Emission Tomography (SPECT) and Positron Emission Tomography (PET) measure the metabolic activity of the brain. The advantage of EEG and MEG over the other brain investigation methods is their high temporal resolution and therefore the possibility to study the brain function in millisecond time scale. Their disadvantage is a poor spatial resolution that limits the possibility to find an exact source location. It was long believed that MEG offers better spatial resolution than EEG due to the high skull resistivity. Nowadays it has been proven that the skull resistivity is not as high as was thought and that EEG actually offers at least as good spatial resolution as MEG or even better. The EEG recording system is cheaper and smaller than the MEG device and it has lower restrictions for the patient movement. In the EEG measurement the spatial resolution can be increased by using larger number of electrodes and low noise measurement instruments. The advantages of EEG measurement and possibilities of better signal quality make it an interesting research topic.
Nowadays due to the development of electronics and especially integrated circuits the possibilities to design small sized active electrodes exist. Active electrodes are a developed version of commonly known passive electrodes and they have an electronic circuit integrated into the electrode. Active electrodes act as an impedance converter and can also have other properties such as amplification, filtering or impedance detection. Active electrodes are more noise tolerant and do not require skin preparation. The need for active electrodes exists especially in high resolution EEG measurements were large number of electrodes are used to obtain the best possible accuracy and signal quality, and in emergency medicine where fast electrode placement is essential. This Master of Science thesis is part of an ongoing research project where objectives are to develop methods and instrumentation to record electric fields of the brain more accurately. The main purpose of this project was to develop a simple and low noise active electrode that can be used in high resolution EEG measurements. Also solutions for electrode wiring were discussed during the design process.
Effects of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin Exposure on Bone Material Properties from Macrostructure to Nanostructure
Finnilä Mikko, Oulun yliopisto
Waste management, gas engines, geological processes and even tobacco smoking may produced aromatic chlorinated compounds called dioxins. Dioxins are notorious compounds, which are very persistent and tend to accumulate in food chain. These compounds are lipophilic and very stable in nature and metabolism. Dioxins are potent endocrine disrupting agents and as such they may also have adverse effects in bone quality as shown in animal experiments. Dioxins have been shown to decrease bone strength, architecture and mineral density. However detailed effects of dioxins on bone material properties are unknown. The aim of this study was to define the effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on bone material properties investigated by nanoindentation and micro-X-ray-diffraction.
The experimental animals were exposed to TCDD in utero and measurements were done in two time points; of post natal days (PND) 35 and 70. Both analysis groups consisted of the animals exposed with the maximal dosage of 1 ìg/kg (N=7-8 per group) or vehicle as a control. Animals were sacrificed and the long bones were dissected out. Peripheral quantitative computed tomography (pQCT) measurements and mechanical testing had been performed previously. The current measurements by nanoindentation and micro-X-ray-diffraction (N=1 per group) were done from the same skeletal locations as the earlier measurements.
Nanoindentation results showed that TCDD delays age-dependent development of bone material (maturation) leaving bones more elastic and softer. Also new findings of correlations between material parameters and bone mineral properties were done.
Cardiovascular Metrics of Autonomic Nervous System Response to Mental Effort
Henelius Andreas, Teknillinen korkeakoulu
The amount of mental workload in occupational tasks has increased, calling for the measurement of mental workload to guarantee safe and efficient working conditions. The measurement of mental workload is especially important in safety-critical professions, where metrics of mental workload could be used as pre-alarm systems for operators. The aim of this thesis was to investigate the response of cardiovascular metrics to mental workload. The correlation between perceived subjective task difficulty, measured using the NASA Task Load Index (TLX), and objective task difficulty, measured by cardiovascular metrics, was investigated. The performance of the Surgical Stress Index (SSI) in the novel application as a metric of mental workload was studied.
Electrocardiograms, continuous blood pressure, respiration and photoplethysmographic (PPG) waveform were recorded from subjects during the performance of a computerised multitask test inducing different levels of mental workload. Several time and frequency-domain cardiovascular metrics were calculated from the recorded signals. Differences between low and high mental workload was studied and the classification performance of the cardiovascular metrics was analysed using prediction probability and receiver operating characteristics (ROC) analysis.
The agreement between subjective and objective workload measures was good. The time-domain metrics performed better than the frequencydomain metrics. The cardiovascular metrics of blood pressure, heart rate and vasoconstriction that best classified mental workload were, in order, average mean blood pressure, average interbeat interval length and standard deviation of photoplethysmographic amplitude. The SSI was found to be a promising index of mental workload.
The findings indicate that the cardiovascular metrics studied in this thesis can be used in measuring mental workload, although further research is needed to establish normal levels for different types and degrees of workload.
Detection of characteristic texture parameters in breast MRI
Holli Kirsi, Tampereen teknillinen yliopisto
Breast cancer is the most common cancer in women. Breast MRI (BMRI) has emerged as a promising technique for detecting, diagnosing, and staging the condition. Automated image analysis aims to extract relevant information from MR images of the breast and improve the accuracy and consistency of image interpretation. Texture analysis (TA) is one possible means of detecting tissue features in biomedical images. The primary aim of this work was to develop a method to easily and efficiently perform the comparison and evaluation of calculated breast MRI texture parameters. The specific aim was to evaluate the parameters that identify the most important breast cancer characteristics and to assess the ability of MRI TA to characterize breast cancer tissue.
Eight patients with histopathologically proven breast cancer were selected in this preliminary study. The texture analysis was performed with MaZda texture application by using histogram, gradient, run-length matrix, co-occurrence matrix, autoregressive model and wavelet parameters for tissue classification. The most discriminant texture features identified by Fisher coefficients and POE+ACC (probability of classification error and average correlation coefficients) between breast cancer tissue and reference tissue from the healthy breast and tissue next to cancer area, between patient, different image series and histological types of (ductal vs. lobular) carcinomas were evaluated. Raw data analysis (RDA), principal component analysis (PCA), linear discriminant analysis (LDA), and nonlinear discriminant analysis (NDA) were run for each subset of images and chosen texture features.
The developed Excel documentation system enabled faster and more effective way to compare different sets of texture features which speeded up the feature comparison process considerably. The results revealed differences in the textures in every imaging series when non-cancer tissue and cancer tissue were compared. Also selected texture parameters were different between the two histological groups. Data analyses RDA, PCA, LDA and NDA performed well on the task discriminating breast cancer area from healthy breast area and the best discrimination results were obtained within two dynamic contrast-enhanced MRI subtraction series.
Texture analysis applied to breast MRI seems to be a potential tool in detecting and discriminating malignant and normal breast tissues and different histopathological breast cancer types, but further research is needed on larger data sets. More studies of textural features in various scales and configurations are necessary in order to select a useful set of texture descriptors, optimal for the specific task of normal and abnormal breast tissue characterization. A part of this work was accepted in the 4th European Biomedical Engineering Congress which took place in November 2008 in Antwerp Belgium. The abstract has been published in the Abstract book of the MBEC 2008 Congress and a short paper in IFMBE Proceedings ECIFMBE 2008 "4th European Conference of the International Federation for Medical and Biological Engineering" Vol. 22, 2008, Antwerp, Belgium.
Design of a micro volume fluorescence measurement device for immunoassays, using up-converting phosphor labels
Koho Sami, Turun yliopisto
Immunoassays are popular methods for monitoring bio-affinity binding processes taking place for example in human blood as a result of an immunoresponse. Both information about the quantity of bound molecules, and about the quality of the binding process, such as speed or specificity can be extracted. After the first immunoassay method presented by Berson and Yallow in 1959, the ability to monitor bio-affinity binding has been found useful in a wide array of fields, from food industry and pharmaceutical development to various bio-sciences and of course medical diagnostics. The need for cost savings particularly in industrial applications, where a high amount of samples needs to be measured daily, has led to the development of automated immunoassay measurement technologies, that would ideally have as short as possible measurement time, with as low as possible material costs. To face these technological challenges TPX measurement technology was developed in the Laboratory of Biophysics at the University of Turku during the late 1990's. Taking advantage of the research group's expertise in optics and luminescent materials, the TPX technology performs micro-volume optical measurements on luminescent labels, taking advantage of a non-linear phenomenon called two-photon up-conversion, seen in certain luminescent labels when excited with high intensity infra-red light.
During the current project a new technology, one might say a further development of the TPX, was realised. The same idea of a micro-volume optical measurement set-up was adapted to use a new form of luminescent probes, called up-converting phosphors (UPT). The new labels have a number of advantages when compared to the ones used for two-photon up-conversion, amongst which degrees of magnitude higher conversion efficiency and the complete absence of photo-bleaching, which at least in theory should allow cost savings to be made in high throughput environments, by way of further reduction in reaction volume. While the basic idea remained the same as in the original TPX, in practice the whole measurement system was re-designed from the beginning. A new embedded control system was realised using an Atmel embedded microcontroller, and new data acquisition circuits for the photo-detectors were realised due to changes in the optical signal. Embedded software for the control logic was written in C programming language, and a PC - software for performing the measurements was realised with LabView. For the optical system, a new continuous wave infra-red laser with 980nm wavelength was acquired, whereas the rest was kept mostly as it was in the original TPX device.
The designed device has been found to be able to detect fluorescence at extremely low excitation light intensities, reaching down to parts of a milliwatt of optical power, whereas for example with rhodamine labels widely used in TPX, at least some tens of milliwatts were required. As in practice the excitation intensity would be kept at reasonably high levels for particle trapping purposes, the superior conversion efficiency of the used labels has been seen to produce a fair amount of signal at low label concentrations. This is all as expected, and would suggest that the designed system performs as it should. However no real assay measurements have been done with the device yet, as the focus has been on the hardware testing. The project is still underway, and the assay performance results will be presented shortly (hopefully already at the LFT days in Kuopio).
An Assessment Tool for Assistive Technology
Lahtinen Ella, Tampereen teknillinen yliopisto
The primary aim of this research was to develop an instrument for the assessment of assistive technology (AT) devices and services. Specifically, the aim was to develop a customer orientated instrument that can be used to assess the user satisfaction with AT.
The research consists of two parts. In the literature study part a number of available, already existing, AT assessment tools are reviewed. In the second part a new assessment tool, the ITSE-Assessment Model, is developed.
ITSE-Assessment Model is a questionnaire that can be filled in by interviewing a final user of AT. The questionnaire consists of four dimensions of AT: usability, utility, quality of service and costs, and it can be used to assess any kinds of AT. The interviewee need only be familiar with the AT that is being assessed, and the interviewer with the ITSE-Assessment Model.
When the ITSE-Assessment Model was developed, some of the functionally perceived instruments presented earlier in this work were used as a background. In addition to this the ideas and opinions of AT specialists were taken into account when determining the content of the questionnaire.
The testing of the ITSE-Assessment Model was carried out, and as a result the ITSE-Assessment Model proved to be a functional instrument that can be used to assess AT closely. However, for easier use of the ITSE-Assessment Model, some improvement recommendations should be carried out and in the future the validity of the instrument should be tested in a more systematic way.
Anyhow, the potential of the ITSE-Assessment Model lies precisely in the examining manner of it. The ITSE-Assessment Model is intended to create discussion about the challenges of AT and to set out improvement proposals. At its best, the ITSE-Assessment Model helps the communication of final users and designers of AT and in this way contributes to the improvement of AT devices and services.
Enhancement of MRI with SQUID arrays and by polarization encoding
Nieminen Jaakko, Teknillinen korkeakoulu
Magnetic resonance imaging (MRI) is a method to study the interior structure of matter. It is based on detecting precession signals from a magnetized sample. The detection can be speeded up by using sensor arrays. In low-field MRI, a sample is polarized in a magnetic field of several millitesla. The polarization is followed by SQUID-based signal detection at microtesla-region fields. In this study, a new encoding method to reduce imaging times in low-field MRI is developed.
In MRI, signals of a sensor array can be written in vector form s(t) = Am(t), where A is a lead field matrix and m(t) contains the components of the voxel magnetizations. In polarization encoding, various polarizing fields affect the initial magnetization of the sample in consecutive measurements. Assume that for the kth measurement the magnetizations are m_k(t) = C_km(t), where C_k is a conversion matrix. Then s_k(t) = AC_km(t). A large signal vector s’ and a generalized lead field matrix A’ are constructed combining s_k:s and AC_k:s row-wise, respectively. Then, s’(t) = A’m(t). If the polarizing fields are chosen properly, rank A’ > rank A, i.e., the number of linearly independent sensors is increased.
The proposed method was tested with low-field MRI simulations using the geometry of a 304 channel SQUID system. Noisy signals from a 15×15 voxels phantom were simulated to the SQUIDs. The inverse problem of image reconstruction was solved using truncated singular value decomposition. The results show that by increasing the number of polarizing fields, the imaging times are reduced. Equally, in a given time the imaging quality can be improved by increasing the number of polarizing fields.
In this study, a new encoding method was developed for MRI. Polarization encoding introduces additional independent sensors and increases the information about the sample. The method reduces imaging times. It is possible to combine the method with other encoding methods; for instance, one dimension can be encoded by polarization encoding and the others by Fourier encoding techniques. Polarization encoding is especially suitable for low-field MRI, because at low fields it is possible to construct sets of polarizing fields. Moreover, at low fields, imaging processes need improvements because current methods are slow and suffer from poor image quality. In addition to MRI, polarization encoding can also be used to improve other imaging techniques, such as magnetorelaxometry.
Unraveling lipoprotein structure via molecular dynamics
Vuorela Timo, Tampereen teknillinen yliopisto
Cardiovascular diseases are the primary cause of death in the industrialized countries. Heart strokes and cerebral infarctions are the end conditions of a cardiovascular disease called atherosclerosis. According to the current view, it is inflicted and further restrained by lipoproteins circulating in blood. In particular, low density lipoprotein (LDL) levels have been found to correlate positively and high density lipoprotein (HDL) levels inversely with the risk of atherosclerosis.
Lipoproteins are complex macromolecular structures designed to carry cholesterol in an esterified form inside the body. They consist of a spherical lipid part surrounded partly by a protein part. Although molecular compositions are quite well understood, the molecular-scale structures have not been solved. In this study we use new computational methods to get information about the molecular-scale ordering in different sized lipoprotein particles.
In this study the structure of two different lipoprotein particles were explored by using classical molecular dynamics methods. The smaller system was constructed to model the lipid part of an HDL particle and the larger system the lipid part of an LDL particle. The LDL particle contains over 3000 molecules and is more than three times larger than the systems in current state-of-the-art studies. Also the time scales in this study are over a magnitude longer than in the earlier studies. To achieve the length and time scales used in this study a method called coarse graining has been used. In coarse graining on average four carbon atoms are substituted by one bead. Coarse graining reduces the complexity of the systems while preserving the behavior of different molecules. Extensive structural analysis is performed for the general structure as well as conformations of each molecule type. Also the dynamics of the whole lipoprotein particles as well as the individual components is studied.
The current view of the lipoprotein structure is that the particles are composed of a disordered hydrophobic core and a hydrophilic surface. The results of this study support the view, but also reveal that there is an additional layer between the core and the surface. The results show significant ordering of the core lipids close to the surface that is not observed in the core region. Cholesteryl esther molecules in the interface order parallel to the phospholipid tails similar to cholesterol molecules. Also triacylglycerols in the interface have more compact packing and have a tendency to orientate in a way that minimizes contacts with water. The diffusion rates of different components are in good agreement with the results from earlier experimental and computational studies.
This study demonstrates that molecular dynamics and especially coarse grained methods can be used to give valuable insight into the structures of lipoproteins, unreachable to other methods. The results are in line with the current view of the lipoprotein particle structure and, in addition, support a recently presented new three-layer structure for lipoprotein particles. This study gives a solid foundation for future projects which aim to better understanding about the structure-function relationship of lipoproteins.