What Can fMRI Tell Us About Mental Illness?

By Neuroskeptic | January 14, 2017 10:53 am

A remarkable and troubling new paper: Addressing reverse inference in psychiatric neuroimaging: Meta-analyses of task-related brain activation in common mental disorders

Icahn School of Medicine researchers Emma Sprooten and colleagues carried out an ambitious task: to pull together the results of every fMRI study which has compared task-related brain activation in people with a mental illness and healthy controls.

Sprooten et al.’s analysis included 537 studies with a total of 21,427 participants. Five mental illnesses were examined: schizophrenia, bipolar disorder, major depressive disorder, anxiety disorders, and obsessive compulsive disorder (OCD). The tasks used in the studies varied and were grouped into four categories: “cognitive”, “social”, “negative valence” and “positive valence”. Only studies which reported at least one group difference in brain activity (“blob“) were included.

The results were rather surprising. It turned out that there were very few differences between the different disorders in terms of the distribution of the group differences across the brain:

Overall, there was no significant effect of diagnosis on the spatial distribution of the reported case-control differences (χ2=232, P=0.27)… for each pair of diagnoses, correlation coefficients across regions ranged between 0.42 and 0.82 and were highly significant (all P<0.001).

In other words, there was little or no diagnostic specificity in the fMRI results. Differences between patients and controls were seen in the same brain regions, regardless of the patients’ diagnosis. While blobs were found in almost every brain region by at least one study, here’s where the differences were most common:


It’s notable that almost all of these top-10 regions fall within the “limbic system” of the brain, i.e. areas which are traditionally thought to be involved in emotion. By contrast, the rest of the cerebral cortex is hardly represented at all.

So mentally ill people have abnormal limbic system responses to various tasks, irrespective of diagnosis. While it’s hardly surprising that an emotional disorder such as anxiety would be linked to limbic dysfunction, it is remarkable that schizophrenia, an entirely different and ostensibly “non-emotional” syndrome, has the same pattern of results.

However, if we look at the data more closely, we can see that there are some differences between the disorders. The differences are not statistically significant, but this might reflect a lack of statistical power:


Schizophrenia (green) was associated with the fewest limbic blobs, so perhaps it is not all that similar to the other disorders.

Still, Sprooten et al. conclude that:

The abnormalities in brain networks and network-regions we can observe with fMRI reflect disorder-general conditions that facilitate the emergence and persistence of symptoms but are insufficient for explaining symptomatic variability across disorders.

This is a pretty radical conclusion because it is implies that most fMRI studies of, say, depression, have not told us anything specific about depression. Instead, the group differences seen in these studies have been reflections of some general ‘transdiagnostic’ trait which is common to depression, anxiety, bipolar disorder etc.

What could this trait be? Sprooten et al. suggest that “the disorders examined here arise from largely overlapping neural network dysfunction”, in other words that the transdiagnostic trait is a neurobiological part of the cause of the various different disorders. But it seems to me that there’s no reason to assume this.

What if the common factor is more straightforward: something like anxiety or stress during the MRI scan?

It’s plausible that patients with mental illness would be more anxious, on average, than healthy controls, especially during an MRI scan which can be a claustrophobic, noisy and stressful experience. This anxiety could well manifest as an altered pattern of task-related brain activity, but this wouldn’t mean that anxiety or anxiety-related neural activity was the cause of any of the disorders. Even increased head movement in the patients could be driving some of these results, although I doubt it can account for all of them.

ResearchBlogging.orgSprooten E, Rasgon A, Goodman M, Carlin A, Leibu E, Lee WH, & Frangou S (2017). Addressing reverse inference in psychiatric neuroimaging: Meta-analyses of task-related brain activation in common mental disorders. Human Brain Mapping PMID: 28067006

  • Bill Flack

    Or maybe most of the people with psychiatric diagnoses have histories of psychological trauma? That might explain the emotional dysregulation. It also argues, if true, for a focus on trauma prevention, imho.

    • rbrto

      and in addition to a possible history of psychological trauma, people with mental illness live under much more stressful/disadvantaged conditions than the rest. Even the life expectancy of people with mental illness is reduced (for reasons not directly related to the disorders)

  • Pingback: What Can fMRI Tell Us About Mental Illness? – Discover Magazine (blog) | EcoGreenData()

  • Pingback: What Can fMRI Tell Us About Mental Illness? - Neuroskeptic()

  • Keith R Laws

    Re anxiety in scanner, I made a similar argument about the psylocybin scanning work a couple of years back http://keithsneuroblog.blogspot.co.uk/2012/03/anxiety-tripping.html?m=1

    • Maia

      Yes, cannot ignore the stress of test equipment/setting re: any experiment.

  • Pingback: What Can fMRI Tell Us About Mental Illness? – Discover Magazine (blog) | APG Editorial()

  • Pingback: What Can fMRI Tell Us About Mental Illness? – Discover Magazine (blog)()

  • Pingback: What Can fMRI Tell Us About Mental Illness? - N...()

  • Andy Anderson

    Scans are used all the time to support biological psychiatry, but unusual brain controls around people who already have been diagnosed with mental illnesses and probably taking medication to relieve their symptoms does not help this study. The study really does not provide a great understanding of the mental illness and brain activity since they do not explain the effects of the medications and brain activity. The study did provide several tasks for participants to perform to show brain activity but they do not examine the current feelings, mood, anxiety, stress, and so on of the participants. Overall, I think it would be great to be able to scan the brain and diagnosis mental illness among other disorders and issues but the scan is not the answer for everything. Too many times people trust what a scan states but think about diagnosing a concussion a scan, cannot tell show you anything but people depend on it to tell them they have one. I think it is a great idea but I think there should be more done to help shine a light on mental illness and challenging the diagnosis and asking “what more can we do,” I think that is the question we need to be asking to help mental illnesses.

  • RMacCoun

    I wonder if perhaps the title here should be a more cautious “What Can fMRI Tell Us About Psychiatric Diagnoses?” rather than “What Can fMRI Tell Us About Mental Illness?”

  • djlewis

    On top of this study — https://www.ncbi.nlm.nih.gov/pubmed/27829086 — which says that neuroimaging cannot characterize depression, it seems that neuroscience has cast exactly zero light on mental disorder and there is no meaningful field we could call “neuropsychiatry”.

    Worse, after 20 intensive years of trying, it appears there is no scientific basis for optimism that neuroscience will succeed soon — or ever!

    Radical conclusion? Yes, but where is the evidence for the contrary?

  • Carolyn Wilshire

    Need to bear in mind that this metanalysis averaged across multiple, incredibly diverse tasks. From verbal fluency to imagery generation to recognition of facial expressions, and many more. Not likely these will activate the same cortical regions, is it? So the averaging they perform has factored out all the specific task effects, and what you are left with is non-task-specific effects. This study tells us what is likely to happen if you average enough group differences across enough widely different tasks. Not sure it tells us much more.

    • ferkan

      I’d actually expect it to show some differences based UPON the fact that we tend to use different paradigms in different patient groups. We’re I suspect more likely to use affective tasks in disorders seen as more affective… eg depression as opposed to schizophrenia.

  • CL

    Meta-analyses are always difficult to draw conclusions from, as a lot of de finer details of group contrasts get grouped in to one. Also, the comorbidity between the above psychiatric disorders is quite high. It may be a bit simplistic to expect one region to be specific for a complex trait, rather, it may be alterations in neurochemistry that pushes the behavior towards extremes in natural variations. That´s why drugs work. And why stroke patients doesn’t typically have clear cut psychiatric manifestations. The neurochemical changes may impact BOLD fMRI, but we don´t know how.
    That said, add insult to injury, the regions that appear also have the highest noise


    and, related, these limbic regions have the lowest ret-retest reliability, with ICCr rarely reaching 0.5 for a relatively solid paradigm like resting state connectivity:

    Unfortunately, I think there is a lot of p-hacking done to find regions like the amygdala in order to tell a coherent and appealing story.

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      “Unfortunately, I think there is a lot of p-hacking done to find regions like the amygdala in order to tell a coherent and appealing story.”

      I agree, and at first I thought that p-hacking and publication bias alone might explain Sprooten et al.’s results i.e. researchers expect to find differences between patients and controls in limbic regions, so this is where the majority of published results are found.

      However, this can’t explain all of the data. I doubt many people would expect to see patient-control differences in the frontal operculum or parahippocampal gyrus, yet they were among the top 10.

      Many cortical regions are important in theories of depression and anxiety and they don’t appear in the top 10 (although it would be nice to see, say, the top 30).

      • ferkan

        Thinking out loud….

        I wonder about other factors. Is it possible that some of these areas are somehow more vulnerable to false positives? (I can’t think of a mechanism right now).

        Or could it be that it’s easier to register these areas accurately than, for instance, parts of the frontal cortex, so it’s simply easier to detect ‘real’ differences (by essentially reducing noise).

        To some extent, the possibility of finding difference must also depend on the fMRI task utilised. If you’re going to use a mood induction paradigm, you’re going to be more likely to find limbic results – although we might actually expect this to drive differences between diagnoses.

        At a first glance, I am struck most by the lack of the parahippocampal gyrus in the OCD data – which I guess fits with the lower ocd-other diagnosis correlations.

        Now assuming that all this data is a real reflection of the underlying dysfunction… I don’t think we should be at all surprised if limbic dysfunction underlies all the disorders examined. I don’t think any of us should consider schizophrenia a non-emotional syndrome. In my experience, in relatively well (not currently psychotic) patients, anxiety and depression are extremely common – and of course, paranoia is a disorder of anxiety/fear as much as anything else.

        I’m more surprised that the amygdala does not really turn up in schizophrenia (figure 2)! And I wonder if this is more a function of the paradigms used.

  • OWilson

    I’d hate to see psychological examination reduced to a Rorschach Test with the final diagnosis left up to a “consensus” :)

    (This may be OK with political or global warming viewpoints, but it still sounds a little totalitarian. Consensus can be famously wrong!)

    • ferkan

      Psychological diagnosis has long been a matter of consensus!

  • http://bpdtransformation.wordpress.com Edward Dantes

    I thought many psychiatric researchers had given up on researching these DSM disorders after Insel, Hyman, Kupfer, Frances et al had publicly admitted that lumping distress into categorical boxes like “schizophrenia” and “major depression” lack validity?

    These labels are just that, assumptions about divisions of experience that might not exist cleanly, “carved at the joints” in nature. There are no known diagnosis-useful biomarkers yet for these labels, as far as I’m aware, although the NIMH won’t stop trying.

    Here are what NIMH and APA leaders said in the last few years

    In 2013, discussing psychiatric diagnosis, the psychiatrist and former National Institute of Mental Health director Steven Hyman stated:

    “The underlying science remains immature…The molecular and cellular underpinnings of psychiatric disorders remain unknown… psychiatric diagnoses seem arbitrary and lack objective tests; and there are no validated biomarkers with which to judge the success of clinical trials.” (emphasis mine)

    Hyman went on to call the DSM model of diagnosis, which includes labels like “schizophrenia,” “Totally wrong… an absolute scientific nightmare.”

    Hyman’s successor at NIMH, psychiatrist Thomas Insel, followed up this criticism by saying:

    “At best, [the DSM is] a dictionary, creating a set of labels and defining each. The weakness is its lack of validity. Unlike our definitions of ischemic heart disease, lymphoma, or AIDS, the DSM diagnoses are based on a consensus about clusters of clinical symptoms, not any objective laboratory measure. In the rest of medicine, this would be equivalent to creating diagnostic systems based on the nature of chest pain or the quality of fever.” (emphasis mine)

    David Kupfer, the DSM 5 chair, while trying to defend the new DSM, admitted that the discovery of biomarkers for supposed illnesses like schizophrenia remains “(D)isappointingly distant… unable to serve us in the here and now.”

    And former DSM-IV head Allen Frances went so far as saying, “There is no definition of a mental disorder. It’s bullshit… these concepts are virtually impossible to define precisely.”

  • practiCalfMRI

    A twitter thread on motion as a candidate explanation:



    Would have started discussion here but been on the road. Sorry!

  • http://brain4biz.wordpress.com/ Brain Molecule Marketing

    This topic is a mess – mainly because of the “top-down” attempt to force fit cultural beliefs and tropes onto the brain data. “Rethinking the Emotional Brain” is a good place to start Joe LeDoux is trying to unwrap the cultural crap and the medical science. It’s hard. He recent papers are good.

    Joe strikes out against “high order concepts” and certainly mental health topics are based on those.

    “Blobology” or the functional specificity of areas is also being debunk by people like Paul Cisek. Hie research shows a massively and instantly parallel processing of behavior all over the brain – effectively – so the model of seria processing by specific areas is pretty old fashioned.

  • http://brain4biz.wordpress.com/ Brain Molecule Marketing

    For the brain geeks among us let me recommend the podcast “Neuroscientists talk shop” and NIH video lectures for some of the best and most evidence-based brain science stuff.

  • Pingback: 011817 | Castaway on B612()

  • M Peirce

    There is a rather simple reverse cause hypothesis available here. Having a mental disorder causes problems for the person who has it. Such people will naturally be concerned with their life prospects, which, as an emotion/mood, is likely to show up across the cohort population as amygdala activation that differs from controls. Because a common effect explanation is available, we lack the warrant to infer that there is any common cause.

  • Pingback: Signal Conduction – 20 January 2017 | Cognitales()

  • jgas

    To the extent that stress is supposed to be a factor in exacerbating all of the conditions,this common stress vulnerability is what is mostly being highlighted as the important factor when compared to controls.
    This commonality between conditions has been noted and theories have existed for as long as attempts at diagnosis.

    It is interesting as well that something like anti-NMDA encephalitis has been observed (by Dalmau) to go through stages of symptoms and recovery involves the same stages BUT in reverse order. The trigger for hypothesising encephalitis as opposed to many and varied psychiatric disorders is usually observed seizure activity. I believe this adds weight to Bartolome Llopis’ theory of a continuum of disturbance of consciousness underlying all pathologies, particular symptomatology being an expression of where you spend most of your time on that continuum.

    Others see mental illness as degrees of loss of complexity:

    Also this

    ‘Psychiatric diagnoses: the weak component of modern research’


    …can’t resist the author’s surname being ‘Angst’ – and psychiatry itself being the argument of the Ur-Woody Allen film :-/

  • Pingback: What Can fMRI Tell Us About Mental Illness? | Circorp()

  • Olmy Olm

    I’m trying to figure out what exactly the differences between the “patients” and the “normals” could be. One thing comes to mind: A recent study showing people living in areas with less green and blue experience more anxiety and other problems than people in areas with lots of green and blue: “Our analyses do indeed show a relationship between GSA [green space availability] and any anxiety disorder, with prevalence being lower when more green space is available…BSA [blue space availability] was related to all health variables, with the exception of any substance use. Moreover, its associations with these health variables were stronger than those of GSA.” https://www.madinamerica.com/2016/12/access-green-blue-spaces-may-improve-mental-health/

    Maybe the “patient” group has a disproportionately large number of city-dwellers? I don’t know, it’s just a wild guess. Is there any data on whether people who live in cities are more likely to encounter psychiatry than others?

    • ferkan

      Unlikely I’d say. Some of the studies for instance will be from London, in which both patients and controls are generally sourced from the same area – normally close to a large university and hospital. We generally try to keep our patients and controls as well matched as possible, and we don’t like to search too far for reasons of practicality. We will source outside the area when we need something specific (for instance twins or people with a rare genetic disorder).

  • Pingback: What Can fMRI Tell Us About Mental Illness? – Welcome to my online library! Your one-stop source for information and entertainment!()

  • Pingback: All mental disorders are brain disorders … not. – Eiko Fried()



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar