An introduction to clinical functional magnetic resonance imaging of the brain

Abstract

Abstract The first papers on functional magnetic resonance imaging (FMRI) were published just fifteen years ago, and in the relative short time since, FMRI has assumed a major role in mapping human brain functions. A search of the term “FMRI” in the medline database reveals that more than 5000 peer reviewed original articles have been produced world-wide, a number which has increased exponentially, since last year alone more than one FMRI article a day was published. Almost nine out of ten publications cover either technical aspects of FMRI or the neurophysiological research concerned with the function of the human brain in normal subjects. The latter field is undergoing an explosive growth due to several distinctly advantageous characteristics of FMRI. These include total noninvasiveness, relatively high spatial and temporal resolution, and the ease of imaging the underlying anatomy. In addition, state-of-the-art whole brain FMRI can now be performed on almost all recently installed clinical MR scanners, and online analysis tools allow the visualization of neuronal activity at the level of the single subject (or patient) in real time. The non-invasive character means that subjects can be studied repeatedly, without harm, allowing for longitudinal studies. These benefits have thus led to a considerable growth of neurophysiological data of the human brain — normal and pathological. The development of applications of FMRI for diagnostic and therapeutic purposes in patients with neurological, neurosurgical, or psychiatric diseases forms about ten per cent of the total number of publications since the beginning of the FMRI era, and this field is growing almost as fast as that of ‘neurophysiological FMRI’. This (at first view) limited share of clinical FMRI is actually a tremendous achievement, considering the fact that clinical FMRI is much more difficult to perform than non-clinical FMRI. The reasons for this are multiple and pertain to all aspects of functional imaging. First, at the design stage of a FMRI experiment, compared to the paradigms or stimuli employed in the neurophysiological study of volunteers, those that will be used to elucidate neuronal activity in patients will have to be adapted to their pathology. Second, at the stage of acquiring the functional scans, acquisition time will often be much more limited than in young healthy volunteers. Third, the post-processing of the FMRI data will often have to be altered or improved due to the presence of ‘unexpected’ signals from the brain abnormalities.