Functional MRI

Functional MRI plays a dominant role amongst brain mapping techniques, in large part due to its non-invasiveness, relatively high spatiotemporal resolution, and the wide availability of clinical MRI scanners. It is important to note that fMRI measures a surrogate signal originating from the vascular response to neuronal mass activity and is therefore subjected to several constraints. Nevertheless, its capacity to map the entire network of brain areas engaged by a specific task is unrivalled among non-invasive techniques like MEG and EEG. The possibility to obtain both functional and anatomical images in the same study session is a further attraction of fMRI.The major impact of the fMRI techniques has been in the field of basic cognitive neuroscience, but they have gained a growing role in translational medicine and clinical practice. Several approaches have been developed for MR imaging of brain function, including contrast agent-enhanced imaging, arterial spin labelling (ASL) and blood oxygenation level dependent (BOLD) imaging, the latter being the most commonly used. Functional activation of the brain can be detected using these techniques through their ability to measure or depict changes in tissue perfusion, blood-volume, or the vascular concentration of deoxy-hemoglobin. While the BOLD technique can provide reliable information on the neuroanatomy underlying transient sensorimotor and to a lesser extent cognitive functions, perfusion techniques are more suitable for studying relatively long term effects on cerebral blood flow (CBF) both at rest or during brain activation. Behind the functional images obtained through BOLD or ASL contrasts, there are complex physical and biophysical principles as well as engineering procedures. The continuous technical improvement of fMRI techniques is most evident in the development of high-field scanner and the optimization of pulse sequences. As with data acquisition, many methods have been proposed for analysing fMRI data, and a variety of these are in general use. The overall aim of such analyses is to produce an image (map) identifying the regions which show significant signal change in response to a specific stimulus or task performed by the subject. In the following sections, after a brief introduction to the physiological correlates of neuronal activity, a description of the main principles and techniques of fMRI based on perfusion and BOLD contrasts will be provided. At the end of the chapter, the main processing steps necessary to obtain a brain activation map by statistical analysis of BOLD data are outlined. Since no gold standard exists among the various methods developed for statistically analysing fMRI data, the most commonly adopted approach, the General Linear Model (GLM) is used in this summary.