Functional MRI: The Environment and Technology for Clinical Application

K. R. Thulborn,B. McCurtain,J. Voyvodic,S. Chang,J. Gillen,J. A. Sweeney
DOI: https://doi.org/10.1007/978-88-470-2194-5_4
1996-01-01
Abstract:As the applications of functional magnetic resonance imaging (fMRI) in clinical radiology become more evident, several broad requirements must be recognized. First, whereas diagnostic MRI uses visual inspection of a few dozen images in which the lesion contrast-to-noise ratio is usually high (>15%), fMRI studies use hundreds of images to detect signal changes (1–5%) well below the visual detection threshold. Second, clinically useful information must be available rapidly for efficient case management, thus requiring the enormous data load of fMRI to be reduced to interpretable form in near real time. Third, the fMRI protocol must be performed quickly and routinely by a MR technologist and nurse. Unlike structural MRI examinations in which the patient is asked merely to hold still, greater patient cooperation is required for cognitive tasks in that the paradigm requires active participation. Preparation of the subject is essential, conveniently performed away from the scanner with a scanner simulator that replicates the fMRI protocol. Such prescan preparation removes novelty effects, is inexpensive compared to scanner time and provides a measure of task performance that can be used to tailor the paradigm to the capabilities of the individual subjects. Such prescan preparation has been very useful for pediatric subjects who, in our experience, take considerable coaching for full cooperation. Acquisition of the fMRI data requires equipment that allows stimulus presentation and response monitoring without loss of scanner performance. Documentation of the timing of coincident images, stimuli and responses must be done accurately and automatically. As all fMRI data are processed off-line, automated documentation avoids delayed and inaccurate communication between technologist and data analyst and allows semi-automated processing.
What problem does this paper attempt to address?