COVID-19 Show message

Skip to content

Post by Former NIMH Director Thomas Insel: Brain Scans – Not Quite Ready for Prime Time

By on

Although brain scans as part of a clinical psychiatric workup might be playing in prime time on some TV infomercials, brain imaging experts say we're not quite there yet.

As editor Dr. Robert Freedman notes in last month's American Journal of Psychiatry, after a string of letters on the subject: “Commercialization of a diagnostic test, even if the underlying procedure such as brain imaging or DNA analysis is approved for human use, strongly indicates to physicians and families that the test adds significant new information to guide clinical judgment. We have published this exchange of letters as part of our responsibility to readers to point out when a procedure may lack sufficient evidence to justify its widespread clinical use."1

The procedure in question is SPECT (single photon emission computed tomography). At issue: the use of SPECT scans as part of a psychiatric diagnostic workup by some clinics. They cast themselves as “early adopters"1  in the vanguard of developing clinical applications for an emerging technology. Their critics say it's premature, pointing out the risks of exposing patients, particularly children, to radiation, and to treatments based on shaky interpretations of scans – possibly in lieu of clinically sound treatments.2 Moreover, more powerful, less invasive technologies have often been supplanting SPECT in psychiatric research over the past decade.

All agree that neuroimaging will likely play an important future role in clinical care for mental illness, as it already does in neurological practice. But let's not jump the gun.3 As with commercialization of genetic tests, entrepreneurial zeal capitalizing on scientific advances needs to be tempered by reality checks. The neuroimaging community owes the mental health community some help in sorting out legitimate from questionable claims.4 In fact, this is a goal of NIMH's budding “neuroethics" portfolio – clarifying the difference between research opportunities and ready-for-prime-time clinical tools.

To be sure, tantalizing new findings are increasingly hitting the mark, raising hopes that brain imaging biomarkers of treatment response may soon be within reach.

For example, a report earlier this year described the use of MEG (magnetoencephalography) to diagnose Post Traumatic Stress Disorder (PTSD) with greater than 90% accuracy.5 Recently, NIMH intramural researchers reported that a MEG signal predicted which depressed patients would respond to the fast-acting experimental antidepressant ketamine.6

In June, scientists at Washington University published fMRI (functional magnetic resonance imaging) findings suggesting they'd identified a biomarker target for depression. The marker was increased functional connectivity of a “dorsal nexus," where three brain networks converge.7 The researchers had earlier shown that depressed patients falter at suppressing activity of one of the networks while processing emotional stimuli. Impaired suppression of this “default mode network" appears to interfere with attending to the tasks of daily living -- mainly by taking the brakes off self-focused activity, such as rumination. Evidence suggests that this excess connectivity could emerge in early development, as experience, such as early life stressors, sculpt the default mode network.8

NIMH has just launched the EMBARC study (Establishing Moderators/Mediators for a Biosignature of Antidepressant Response in Clinical Care) in search of clinical, neuroimaging, or physiological measures that will help clinicians tailor treatments for any individual with depression. Other clinical applications of imaging will likely emerge from large studies of normal brain development now underway, such as the Pediatric Magnetic Resonance Imaging (MRI) Study of Normal Brain Development and the Human Connectome Project. Eventually, such efforts will yield definitive benchmarks against which clinicians will be able to compare scan data from patients in their clinics.

As exciting as such advances are, brain imaging is still primarily a research tool when it comes to mental disorders. Scans are appropriate for ruling out obvious pathology, like brain tumors, as possible causes of symptoms. The differences in brain structure and activity seen in disorders like schizophrenia or ADHD, for example, are typically only meaningful when comparing group statistics. There is simply too much individual variation in brain structure and function for an individual's scan to be diagnostic or predictive, given the current state of the science.


1 Brain SPECT imaging in clinical practice. Amen D. Am J Psychiatry. 2010 Sep;167(9):1125; author reply 1125-6. No abstract available. PMID: 20826857

2 Scientifically unfounded claims in diagnosing and treating patients. Adinoff B, Devous M. Am J Psychiatry. 2010 May;167(5):598. No abstract available. PMID: 20439400

3 Book Forum. Leuchter AF. Am J Psychiatry 166:625, May 2009 doi: 10.1176/appi.ajp.2009.08121843

4 A picture is worth a thousand dollars. Farah MJ. J Cogn Neurosci. 2009 Apr;21(4):623-4. No abstract available. PMID: 19296729

5 The synchronous neural interactions test as a functional neuromarker for post-traumatic stress disorder (PTSD): a robust classification method based on the bootstrap. Georgopoulos AP, Tan H-RM, Lewis SM, Leuthold AC, Winskowski AM, Lynch JK, Engdahl B. J Neural Eng. 2010 Feb;7(1):16011. Epub 2010 Jan 20.PMID: 20086271

6 Anterior cingulate desynchronization and functional connectivity with the amygdala during a working memory task predict rapid antidepressant response to ketamine. Salvadore G, Cornwell BR, Sambataro F, Latov D, Colon-Rosario V, Carver F, Holroyd T, DiazGranados N, Machado-Vieira R, Grillon C, Drevets WC, Zarate CA Jr. Neuropsychopharmacology. 2010 Jun;35(7):1415-22. Epub 2010 Mar 10.PMID: 20393460

7 Resting-state functional MRI in depression unmasks increased connectivity between networks via the dorsal nexus. Sheline YI, Price JL, Yan Z, Mintun MA. Proc Natl Acad Sci U S A. 2010 Jun 15;107(24):11020-5. Epub 2010 Jun 1.PMID: 20534464

8 The default mode network and self-referential processes in depression.Sheline YI, Barch DM, Price JL, Rundle MM, Vaishnavi SN, Snyder AZ, Mintun MA, Wang S, Coalson RS, Raichle ME. Proc Natl Acad Sci U S A. 2009 Feb 10;106(6):1942-7. Epub 2009 Jan 26.PMID: 19171889