Investigation of image processing algorithms for medical applications

Abstract

Functional Magnetic Resonance Imaging (fMRI) is a neuroimaging non-invasive method that uses magnetic resonance imaging. It deals with the detection of brain areas, which are usually involved in work or an emotion. It is basically, a way to providing information by measuring brain activity. One of the methods for fMRI is based on Blood Oxygenation Level Dependent (BOLD) signal fluctuation that takes place because of the hemodynamic and metabolic sequelae of neuronal responses. This is widely used. It works by detecting the changes in blood oxygenation levels and flow, which occur because of neural activity i.e. when parts of the brain are more active, then it consumes a higher level of oxygen. To encounter this larger demand, the blood flow increases to the active region. Functional Magnetic Resonance Imaging can be used to generate activation maps, which show parts of the brain that, are concerned with a specific mental process.

There have been a lot of researches that address the human cognitive process, which tries to understand which part of the brain is activated when a person performs a specific task and then switches to another. In this Final Year Project, this analysis is done via a univariate approach. Hence, each particular voxel is treated uniquely. The typical method for fMRI data analysis is the General Linear Model. The author has also used the Support Vector Machine and the Power Spectum analysis for data analysis and generating activation maps.