INTRODUCTION
        Within the past 15 years psychiatry has rediscovered the brain. Like our neurological cousins, we now fully recognize the obvious fact that normal and abnormal behavior is mediated through neuronal function. The reasons for this rediscovery are complex, fueled in part by the revolution in neuropsychopharmacology. In addition, unlike just 15 years ago, we now have a multitude of different tools for directly imaging the living working human brain. When these tools first appeared, many predicted that they would transform psychiatry, substituting a high-tech scan for the skillful clinical exam. Optimism and hyperbole were quite high. We now realize that things are not as simple as we once thought, and that there will be many years of research before brain imaging totally impacts psychiatry. The pendulum swings of optimism and pessimism with respect to imaging are paralleled in the discussions of how the field of genetics will change psychiatry, while in fact there has not been a single replicable gene finding in the major psychiatric disorders.
        Neuroimaging is beginning to live up to its potential in the area of psychiatric research. But what exactly is the current and near future role of brain imaging in the everyday practice of primary psychiatry? This issue is designed to serve as a primer of the basic facts that a psychiatrist should know about neuroimaging, describing the methods, reviewing some of the more important research findings, and with an in-depth discussion of the clinical situations where structural or functional imaging may play a role. In this advancing field, this discussion could change rapidly so readers should be aware of potential new developments.

WHERE WILL IT ALL LEAD?: POTENTIAL INSIGHTS FROM THE EEG
        What will be the eventual role of imaging in psychiatry in the new millennium? Will diagnostic interviews be replaced by scans which make a diagnosis or predict treatment response? A useful analogy may exist with respect to the development of the electroencephalogram (EEG) in the diagnosis and treatment of epilepsy. When the EEG was first discovered in the 1930's, many neurologists assumed that it would transform the study of epilepsy. While the EEG has certainlyimpacted on neurology to some extent, this has not happened in the way that many would have predicted. Over 50 years after the EEG's discovery, the diagnosis of epilepsy remains a clinical diagnosis, based on history and physical exam 1. Although the EEG can be helpful and confirmatory, and even quite specific in some cases such as Creutsfeld Jacob Disease, the diagnosis of epilepsy can be made in the presence of a normal EEG, and conversely, an abnormal EEG does not make the diagnosis. While it is true that most patients with epilepsy undergo an EEG at some stage of their workup, the EEG findings are necessarily incorporated with physical exam and clinical findings to make final diagnostic and treatment recommendations. In clinical epilepsy cases where treatment focuses more on regional brain interventions than pharmacological management, the EEG assumes a greater role. Thus, in patients undergoing temporal lobectomy, proper identification of the seizure source with in-depth EEG studies is quite important.
        Some would argue that functional imaging will perhaps be for psychiatry as the EEG is now for neurology. Imaging studies will always necessarily be integrated with other aspects of the biopsychosocial model for proper understanding in psychiatry. That is, even full understanding of how a psychiatric disease is disordered at the regional level, seen on a brain scan, will not give full understanding of the disease in general or the way the disease manifests in an individual. The information about regional activity will necessarily have to be integrated with other information above and below in the biopsychosocial model.

POTENTIAL PITFALLS
        Another source of potential problems is the attempt to overlocalize brain functions to specific brain regions and to ask neuroimaging to do too much for us. This overlocalization was common in the last century, in the field of phrenology.

Some of the statements about localization of functions to brain regions, using the size and the shape of the skull to predict, now strike us as clearly absurd. However, it is important to remember that we may be committing similar errors when we, for example, state that working memory resides in a certain part of the prefrontal lobe.The brain is likely much more complex than a simple one-to- one relationship between region and function.
 
        In a similar vein, we often ask a new technology to answer questions for us that are difficult to conceptualize. For example, figure 3 is from a paper at the turn of the century where scientists were then excited about the then new tool of photography and x-rays. They hoped that the new technology might shed light on the complex question of why someone becomes a criminal. Skull x-rays were taken from 12 criminals and overlaid,

thus perhaps yielding a composite of the "criminal face." We now think that it is unlikely that facial bones are the answer as to why someone becomes a criminal, or that skull x-rays could be used to predict future truant behavior. These past studies should give us pause when we try to do the same thing with brain imaging and complex behaviors like criminal activity.
 
BACKGROUND INFORMATION : STRUCTURE AND FUNCTION
        With these cautions aside, what are the new imaging tools that are available, and what key concepts are important for using them? One of the most important distinctions in imaging is whether one is looking at brain structure or brain function. As the British Neurologist John Hughlings Jackson noted in the 1880Õs, brain structure does not equal function and vice versa 2. That is, structural brain damage, such as a tumor, can produce either obliteration of the function normally subserved by that portion, or it can heighten the function of that portion of the brain 3. Additionally, one can have normal brain structure (at least to the limit of current technology) and have markedly abnormal function (that is, areas of the brain that are normal structurally, but are "off-line" functionally). In contrast to CT and traditional MRI which images the structure of the brain, several technologies have been developed recently with the power to look at brain function. For example, figure 4 shows a structural MRI scan (grey) in the transverse plane of a healthy 25year old businessman. He was born normally but contracted bacterial meningitis at age 6 months, with an infarct damaging his left temporal lobe (dark areas, where language resides in most people). However, his brain recovered and as an adult he has normal language function. Areas of increased blood flow during a word generation
task are placed in color on the structural scan. Note how in the presence of grossly abnormal structure he has normal function and speaks fluently (and is a college graduate). His brain has moved the function to other regions.
 

WHAT'S BEHIND THE IMAGE - THE SOURCE OF THE SIGNAL
        As will be explained in more detail later, CT and MRI can image brain structure - what the brain would look like if you could remove it from the skull and put it on the table. Radiotracer based techniques such as PET and SPECT provide an image of brain activity or function. Recently the entire area became more complex when scientists figured out how to obtain functional activity with an MRI scanner. Thus, MRI alone can provide images of both structure and function.
        How does one measure brain or neuronal activity? Brain regions that are more active consume more glucose for energy consumption and receive more blood flow in order to provide oxygen and carry off waste. Thus areas that are more active will provide a larger signal if one injects a radiotracer that is coupled to blood flow (oxygen 15 PET) or glucose (FDG PET). Under most conditions, blood flow and metabolism are coupled 4. Using more sophisticated tracers one can label specific neurotransmitter receptors and transporters, providing information about regional pharmacological activity.
        A problem that plagues the entire field of functional neuroimaging is the exact relationship between an increase in regional brain activity and the behavioral task or disease being studied. That is, does the area with increased activity actually cause the behavior or disease, or is it an attempt by the brain to regulate or dampen the behavior or disease, or has the study been poorly designed and is the area only incidentally activated? To answer these issues, one has to review old case reports of what happened when people had strokes or trauma to particular regions, and thus deduce what behaviors fall off if the region is removed or stimulated. Table 1 provides a rough overview of this parcelling out of behaviors to different brain regions. Coupling functional imaging with a new method of non-invasively stimulating the brain (transcranial magnetic stimulation, TMS) may allow one to get around this problem of tightly understanding the link between regional brain activity and behavior ^5,6.
 

DISTINCTION BETWEEN RESEARCH FINDINGS AND CLINICAL USEFULNESS
        Another important point about functional imaging and psychiatry is that while there may be important new research findings in an area, those may not be applicable in a clinical setting in an individual patient. For example, numerous studies have now demonstrated that in OCD patients, there is abnormal activity in the caudate and orbitofrontal regions 7-9. However, this reproducible and important research finding, performed in groups of individuals, does not translate into using imaging in single clinical cases. A functional scan of an OCD patient may or may not show this abnormality.

SUMMARY
        Thus, functional imaging and psychiatry is at a point of development where imaging tools are transforming our concepts of neuropsychiatric diseases and firmly grounding psychiatric pathology back in the brain, but many of the important research findings are not yet able to translated into clinical practice. This issue of Primary Psychiatry obviously cannot review all of the exciting imaging research findings in all psychiatric disorders (for more information see 8,10-13). Instead, we have attempted in the following sections to overview the recent research findings and largely limit discussion to those with immediate or imminent clinical applications.

REFERENCES:
1. Engel J. Seizures and Epilepsy. Philadelphia, PA: FA Davis and Company, 1989.
 
2. Jackson JH. Observations on the localisation of movements in the cerebral hemispheres. W Riding Lun Asylum Med Reports 1873; 3:175-190.
 
3. Jackson JH. On Temporary Mental Disorders After Epileptic Paroxysms. W Riding Lun Asylum Med Reports 1874; 5:103-129.
 
4. Sokoloff L. Relation between physiological function and energy metabolism in the central nervous system. J Neurochem 1977; 29:13-26.
 
5. Bohning DE, Shastri A, Nahas Z, Lorberbaum JP, Anderson SW, Dannels W, et al. Echoplanar BOLD fMRI of Brain Activation Induced by Concurrent Transcranial Magnetic Stimulation (TMS). Investigative Radiology 1998; In press.

 6. Fox P, Ingham R, George MS, Mayberg HS, Ingham J, Roby J, et al. Imaging Human Intra-Cerebral Connectivity by PET During TMS. NeuroReport 1997; 8:2787-2791.
 
7. Baxter LR, Schwartz JM, Guze BH, Bergman K, Szuba MP. PET imaging in obsessive compulsive disorder with and without depression. J Clin Psychiatry 1990; 51:61-69.
 
8. George MS. The Contributions of PET and SPECT Toward a Psychopharmacologic Neuroanatomy of Obsessive-Compulsive Disorder. In: Hindmarch I, Stonier P, editors. Human Psychopharmacology: Measures and Methods, Vol.4. London: John Wiley and Sons, 1993:99-122.
 
9. Rauch SL, Jenike MA, Alpert NM, Baer L, Breiter HCR, Savage CR, et al. Regional Cerebral Blood Flow Measured During Symptom Provocation in Obsessive-Compulsive Disorder Using 15O-labelled CO2 and Positron Emission Tomography.. Arch Gen Psychiatry 1994; 51:62
 
10. George MS, Ring HA, Costa DC, Ell PJ, Kouris K, Jarritt P. Neuroactivation and neuroimaging with SPET. London: Springer-Verlag, 1991.

11. Roland PE. Brain Activation. New York: Wiley-Liss, 1993.
 
12. George MS, Ketter TA, Kimbrell TA, Post RM. Brain Imaging in Mania. In: Goodnick PJ, editor. Mania. Washington, DC: American Psychiatric Press, 1997:191-241.
 
13. George MS, Ketter TA, Post RM. What Functional Imaging Studies Have Revealed About the Brain Basis of Mood and Emotion. In: Panksepp J, editor. Advances in Biological Psychiatry. Greenwich, Conn. JAI Press, 1996:63-113.