A Survey of FMRI Data Analysis Methods

Functional magnetic resonance imaging (fMRI) is a procedure in which the MR imaging device is used to observe genuine or task induced brain activity networks. Finding and measuring the patterns of brain activity determine which part of the brain is handling critical functions. These patterns also monitor the effects of different conditions (i.e. learning new habits, stroke and trauma, and neurodegenerative diseases). Functional MRI measures the physiological fluctuations in brain cells and relies on the fact that neurons with more activity consume more oxygen. The output of an fMRI scan is a series of raw images, meaning they contain errors. The errors are due to thermal noise, system noise, unwanted head movement during the scan, and other sources. Hence, some preprocessing on the data is required to maximize the information that can be obtained from the images. In recent years, machine learning methods have been developed and trained to use fMRI data as input and aid medical professionals for diagnostic purposes. In this paper the preprocessing and several methods of analyzing the fMRI data is discussed.