Work

Numerical methods for improved signal to noise ratios in spatiotemporal biomedical data

Year: 2010
Type: dissertation
Abstract: Magnetocardiography (MCG) is a technique to measure the magnetic fields\nproduced by electrical activity in the heart. The interpretation of MCG\nsignals is difficult because of different disturbances and noise. Several\nmethods have been suggested for noise reduction in MCG data such as\naveraging, pass or stop band filters, and statistical based methods, but a\nunified framework that takes into account different typologies of MCG\nsignals (rest, stress, and patients with an already ICD– Implanted\nCardioverter Defibrillator- implanted) using an adequate number of\nrecordings is still missing. Consequently, the main aim of the thesis is to\ndevelop methods for noise and artifacts treatment. Due to the\nnon-stationarity (NS) of the noise, the conventional ensemble averaging of\nthe data does not yield the theoretical improvement. In order to overcome\nthis problem an average procedure that ignores the noisiest beats is\napplied. The results of this averaging procedure confirms that in case of\nNS, the Signal to Noise Ratio (SNR) does not behave as expected, but\nreaches a maximum after a certain number of selected beats. Furthermore, a\ntheoretical proof of this result is given. The second part of the thesis\ndeals with techniques based on Blind Source Separation (BSS), as\npreprocessing step for the averaging procedure, in case of MCG signals with\nlow SNR. Different BSS algorithms are compared in order to find the best\none in terms of noise reduction, separation, and computational time for\neach data typology. A drawback of BSS techniques is the order of the\nsources that cannot be determined a priori; for this reason 3 methods\n(based on different statistical principles) have been developed for the\nretrieval of cardiac signals. The last part of the thesis deals with the\napplication of BSS methods to a category of signals not yet analyzed:\npatients with ICD implanted. It is shown that it is possible to extract the\ncardiac signal also in such noisy data, although not automatically. The\nTemporal Decorrelation source SEParation (TDSEP) algorithm outperforms the\nother BSS methods. This thesis shows that, applying novel automatic\nroutines for the removal of noise and artifacts, MCG data could be used in\nclinical environments.
Language: English
Field: Medicine
Open Access status: green
OpenAlex ID: W2125836938
Cites: 101
Cited by: 1
Related to: 10