A new patient walked into my psychopharmacology clinic recently. She seemed not unusual at first. A line in the chart noted that she was somewhat wary of psychotropic medications, but then many patients are, and I don't mind that attitude since I also subscribe to the less-is-more theory of psychiatric meds.

She was on a fraction of the normal dose of a common antidepressant, which, she reported, was her only medication. It seemed unlikely that this small dose was having any effect, positive or negative, on her mental health; but her mood was stable and she was sleeping and eating well. Things looked good. The main points covered, I asked if she had any other concerns.

"Well, I had my brain imaged," she offered. Brain imaging? Did she have a neurological disorder I hadn't known of? I asked which physician had ordered the tests.

"Oh, I did it myself. I saw it on TV and thought it would be a good idea."

Brain imaging advertised on TV, direct-to-consumer?

She handed me a sheaf of papers, of which the first few pages read like Baby's First Book of Neuroanatomy crossed with the New Age Guide to Herbal Remedies. "Prefrontal cortex: Planning ! Attention!... L-tyrosine!... Anterior cingulate: Cooperation! Flexibility!... St John's Wort!" There was then a list of "hyperactive" and "hypoactive" regions in this patient’s brain.

She'd paid out of pocket to have her brain SPECTed at a clinic that advertises on late-night TV.

Where to begin.

Let me just state up front that there is no established role for functional brain imaging in psychiatry at this time. None. You can't look at a picture of blood flow (or any other parameter) in the brain and make a diagnosis. (SPECT has some utility in distinguishing Alzheimer's from vascular dementia, but that's more neurology than psychiatry.)

For one thing, despite the existence of a number of studies looking at blood flow in depression, there is not a clear consensus on which areas of the brain are most involved.  So there's no established baseline for where we're supposed to be looking.  People typically pick a bunch of areas that seem relevant and, when they publish, highlight the ones that happened to look most different between the depressed and nondepressed groups.  

 That's fine and appropriate for the research stage where you try to figure out how stuff works, but it's emphatically not fine for patient care, where the idea is to take well-established knowledge and apply it to patients' health.

 Next, we have no standard against which to judge baseline levels of regional activity. Blood demand fluctuates on a moment-to-moment basis depending on what the brain is doing at any particular time. The variation in demand between individuals and times is not very well described.  

In a research study you have a defined control group.  But when you look at one patient's brain, you don't have that baseline.  So my patient's brain regions were over- or underactive compared to what? Somebody else's brain thinking sad thoughts? Her own brain thinking happy thoughts? The average of ten other people's brains doing a crossword puzzle? Any way you slice it, it doesn't sound informative.

Even when looking at more stable and reliable imaging correlates of clinical depression, in which a clear average population difference exists between depressed individuals and normal controls, there is so much overlap between the two groups that you can't usually infer a person's mood from his brain scan alone. Check out this graph from a review by J.H. Meyer, showing MAO-A density in different brain regions.






While there are clear differences between the populations on the whole, it's also just as clear that an isolated brain scan likely won't tell you much about whether that individual is depressed or not. There are just too many people in the overlap region.

As you can tell from these links, science is still in the phase where we use clinical data as the gold standard against which to judge the validity of imaging results. Going the other way round - starting with the scan and inferring the diagnosis - is something that's far away from our current level of understanding.

Consonant with this, the 'interpretation' of this woman's SPECT scan was nothing you couldn't have figured out from talking to her for a few minutes. After citing a number of areas in which 'dysfunction' was discovered by the SPECT scan, it described some related problems she might have, such as "negativity, guilt, blame, irritability"  - all the kinds of things you could diagnose more easily from a cheap interview than from an expensive brain scan.

The recommendations included a number of OTC supplements (without indication of dosage, of course, since these supplements are largely unregulated, little is known about optimal dosing and in any case actual content may vary wildly from what's written on the label). Some of them were items that have some evidence for their utility (e.g., omega-3 fatty acids); others seemed relatively benign (e.g. Coenzyme Q10) but with little available evidence regarding their use in depression; and others (e.g., St John's Wort - see this FDA advisory) can be positively harmful under the wrong circumstances.

My patient was on all of the recommended supplements - some six or seven different pills - despite having declared the Lexapro as her only medication. This is an incredibly frequent error made by patients, who are soothed by the 'supplement' label into thinking the items are somehow safer or less likely to produce side effects than chemically prepared medications. In fact, there are a number of supplements that have produced significant health problems in their users (e.g., ma huang, which was ultimately banned by the FDA), and since their preparation is poorly regulated, both the dose of the medication and the number and identity of the compounds present are pretty much up for grabs. These facts make supplements a rather riskier bet than prescription preparations.

There were some other very general recommendations of the type that any mainstream  practitioner would typically make: cognitive behavioral psychotherapy, good social and emotional support, exercise, self-relaxation, and a balanced diet without excess use of nicotine and caffeine. All useful advice, none of it requiring the service of gamma ray scans.

I fully expect that brain imaging will one day have a place in psychiatric practice. However, that day is not today, and I find it upsetting when my patients are conned into paying good money for a useless procedure, a dose of radiation, and some occasionally inappropriate advice.

I do find it curious that so many people place deeper trust in the uncharted waters of alternative medicine than they do in the well-documented, frequently evidence-based recommendations of the standard medical establishment. Things aren't always peachy in the world of traditional medicine; doctors have done a lot of harm as well as a lot of good. But I like to think that as a profession we ultimately learn from our mistakes, and I'd submit as evidence the enormous strides modern medicine has made in extending the productive lifespan, reducing infant mortality, etc etc etc. And in a plug for my own specialty of psychiatry, although there are many patients who are beyond our present capacity to help, I know many others who have unquestionably been saved from suicide or from personal and social destruction.

When someone shows me a randomized controlled trial that demonstrates the utility of a supplement, I'll be happy to use it and recommend it. (I already do recommend both fish oil and melatonin under the appropriate circumstances.) Until then, it's just a black box; and that's not something I'm comfortable giving to a patient.