Analysis of Magnetic Resonance Imaging

Cognitive processes are widely distributed across the whole brain, involving interacting and overlapping brain regions. Magnetic resonance imaging (MRI) provides a noninvasive, in vivo, quantitative measurement of psycho-physiologically relevant parameters that are related to cognitive operations in the normal and abnormal brain. The combination of sophisticated experimental designs and powerful statistical analysis of MRI signals has become a powerful remarkably useful tool for cognitive neuroscience and psychology. The present chapter tries to give an overview
of the main statistical tools used in structural and functional MRI analysis. In addition, we will also consider the analysis (preprocessing) and treatment of magnetic resonance images. While the following sections are restricted to MRI, certain points discussed below can also be applied to other neuroimaging techniques like positron emission tomography (PET).

While the electroencephalography (EEG) section in Chapter 28 is focused on the temporal properties of the signals, this chapter emphasizes the spatial aspects. This is a reflection of the two major approaches of MRI in cognitive neuroscience: (1) structural imaging is performed to obtain an accurate description of the morphological characteristics of the studied brain; and (2) functional imaging provides information about the average hemodynamic response in each part of the brain (which is compartmentalized into many small volume units, ‘voxels’) when a subject performs a specific task. In both approaches, a detailed (structural or functional) spatial image of the brain is obtained.As the temporal resolution of the hemodynamic response is relatively slow, this has led to a preference for MR analysis to describe spatial aspects of the response.