Considering the complexities of the fMRI method discussed above, how does fMRI fair when judged in light of current standards of validity, reliability, standardization, and normative description? The construct validity of fMRI as ameasure of neural function at the level of brain systems is becoming well established. Considerable knowledge about abnormalities of brain response, measured by fMRI, is also becoming available for many neuropsychiatric conditions. Less is known about the predictive validity of fMRI data. Reliability studies are becoming increasingly more common, although not all statistical models of signal reproducibility are relevant to understanding the repeatability of fMRI measurements for individuals. It is clear, however, that reliability needs to be established not only for each new behavioral paradigm used in an fMRI study but alsowhen significant changes have been made in the core image-acquisition protocol or to the image-analysis path. The myriad ways in which fMRI studies can vary from center to center has complicated the standardization of methods across sites. Little standardization currently exists. However, several multi-site studies already underway are aimed at standardizing fMRI data collection for selected behavioral paradigms. These studies are likely to stimulate further efforts at standardization in the future. Unfortunatley, this lack of standardization has prevented the field from acquiring much normative data, although some individual sites might have acquired sufficient data over the years to provide local norms. The six papers included in this issue provide a detailed introduction to the potential uses of fMRI in clinical practice. Because we believe that the translation of basic and clinical neuroscience into clinical practice does not happen passively, we have asked each author to demonstrate some of the salient points of their reviews through case studies pulled from their research programs and clinical practice. This integration of the general with the particular captures the translation of knowledge that led to the clinical validation of neuropsychological tests, in which the results of group studies were demonstrated in the protocols of individual patients. The first paper in the series is a primer that introduces readers to the basic principles of MRI, discusses the BOLD method, describes magnetic resonance techniques for measuring cerebral blood flow, and finishes with a discussion MR methods to measure changes in oxygen utilization. This paper argues that baseline levels of blood volume and oxygen utilization can have significant effects on the observed BOLD signal. The Wierenga and Bondi paper describes patterns of fMRI abnormality associated with the risk of developing Alzheimer's disease (AD), as well as patterns observed in early AD. The authors conclude that fMRI changes precede symptoms for individuals at risk for AD, laying the foundation for longitudinal studies of the dynamic brain changes that occur as individuals become symptomatic. The two case studies comparing young and old individuals emphasize the importance of accounting for baseline effects when interpreting the BOLD response of individuals. The Bookheimer paper discusses the most common current use of fMRI in clinical practice - presurgical language mapping. This paper discusses the challenge of using fMRI to identify cortex whose surgical excision would lead to permanent language impairment from cortex whose excision would lead to transient or no language impairment. The paper by Crosson and colleagues discusses the exciting use of fMRI to monitor rehabilitation interventions in aphasia. The review aims to identify conditions that predict the involvement of left hemisphere areas in the recovery from aphasia and conditions that predict right hemispheric involvement. Once these determinants are known, aphasia treatment could be tailored to the brain systems most likely to participate in language recovery. The Paulus and Stein paper reviews the growing use of fMRI in drug discovery. The paper emphasizes the need to develop standardized behavioral probes appropriate for drug research. The final paper, by Bobholz and colleagues, discusses the newly-available clinical practice codes used to bill insurers for fMRI evaluations. Their discussion leads to questions about how individuals should be trained to perform fMRI studies and foreshadows future discussions about credentialing procedures. Training strategies for fMRI clinicians will differ depending on whether the multidisciplinary team model or the expert clinician model becomes the dominant approach to fMRI practice. The multidisciplinary model can be implemented rapidly at institutions where the requisite disciplines are already present. Yet, not all institutions can support the expertise in behavioral sciences, functional neuroanatomy, MR physics, image processing, psychopathology, neuropathology, and radiology needed for multidisciplinary teams. For such institutions, the expert clinician model is an attractive alternative. Given the fundamental yet growing knowledge about functional brain systems and behavioral analysis required to interpret fMRI findings, the expertise needed by a single clinician will require markedly different training from that found in traditional radiology residencies. Rather, a new clinical field - behavioral neuroradiology - with its own body of knowledge is needed to adequately train clinicians to interpret images of the brain substrate underlying normal and impaired brain function. © Springer Science+Business Media, LLC 2007.
