Aim of the project was to develop a set of automatic tools for processing 1HNMR human brain data. Nuclear Magnetic Resonance (abr. NMR) is a very popular and widely used medical diagnostic technique. It is based on a phenomenon of magnetic resonance that take place for a nucleons placed in a homogenous magnetic field. As a next step series of electromagnetic pulses (abr. EMP) is applied that cause a change of magnetic spin. When pulse is released nucleons returns to the state of equilibrium. As a natural consequence an electromagnetic wave is emitted and measured by a receiver coil in the NMR scanner. According to the structure of a pulse a result might be observed as a wave consists of harmonics emitted from different types of compounds. Such a type of NMR test is called Magnetic Resonance Spectroscopy (abr. MRS) and is widely used in detection of tumour metabolite profile. Raw signal contains number of unwanted components that must be removed from the signal before metabolite quantification. It is done in a step called pre-processing. A step after, called postprocessing, consist of modelling techniques that are used to determine a very accurate value for area under specific peak or group of peaks in frequency spectrum. Author with his supervisor proposed an automatic tool called GNMR for full processing of described data. Second very popular NMR test is called Diffusion Tensor Imaging (abr. DTI). In this procedure EMP is emitted in such a configuration that it is possible to determine a water nucleons flow in number of directions. Raw data obtained from such are pre-processed in order to eliminate noise and bad measurements. In the final step information about nucleon movement for multiple directions is translated into geometrical estimate named tensor. Such sophisticated information is then used in a process of tractography (estimation of neural tracks) or is used for calculation of anisotropy maps that are very useful in tumour diagnostic process. During visit in DKFZ Heidelberg author proposed an implementation of software solution for tensor estimation. Both mentioned solutions are described in details in the paper.
[1]
D. Rubin,et al.
Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper
,
1977
.
[2]
Robin A. de Graaf.
In vivo NMR spectroscopy
,
1998
.
[3]
N. Intrator,et al.
Free water elimination and mapping from diffusion MRI
,
2009,
Magnetic resonance in medicine.
[4]
Ronald E. Goldstein,et al.
Principles and techniques
,
2009
.
[5]
J. Thiran,et al.
Understanding diffusion MR imaging techniques: from scalar diffusion-weighted imaging to diffusion tensor imaging and beyond.
,
2006,
Radiographics : a review publication of the Radiological Society of North America, Inc.
[6]
R. Tarnawski,et al.
Optimizing of MALDI-ToF-based low-molecular-weight serum proteome pattern analysis in detection of breast cancer patients; the effect of albumin removal on classification performance.
,
2010,
Neoplasma.
[7]
D. Tuch.
Q‐ball imaging
,
2004,
Magnetic resonance in medicine.
[8]
Tao Wang,et al.
Automics: an integrated platform for NMR-based metabonomics spectral processing and data analysis
,
2009,
BMC Bioinformatics.
[9]
David M. Rocke,et al.
Baseline Correction for NMR Spectroscopic Metabolomics Data Analysis
,
2008,
BMC Bioinformatics.