The Myth of the Brain’s Pain Matrix?

By Neuroskeptic | January 9, 2016 8:51 am

How does the brain encode physical pain? Which brain areas (if any) respond only to painful stimuli?

A new paper reports that one supposedly “pain-selective” brain region, the posterior insula, doesn’t actually specifically encode pain – it activates in response to diverse non-painful stimuli as well. The study appears in PLoS Biology and it comes from Giulia Liberati and colleagues of the Université catholique de Louvain, Belgium.

Liberati et al. found that the insula responds to non-noxious visual, auditory and tactile stimuli, as well as to painful ones. This was true of every part of the insula they examined, ruling out the possibility of a pain-specific spot within the region.

The study included six patients who had electrodes implanted in the insula as part of the programme of treatment for their epilepsy (intracerebral recording). Liberati et al. recorded local field potentials (LFPs) from a total of 72 different sites across all patients, 47 of which were in the insula. Painful stimuli were provided using a small laser. Here’s the results from one patient, showing that all categories of stimuli evoked responses.


The insula is a region of the cerebral cortex whose function is poorly understood. Among other hypotheses, the anterior portion of the insula has been associated with taste perception and with certain emotions, especially disgust, while the posterior part has been seen as encoding the intensity of pain. Liberati et al.’s data are likely to require a revision of all of those theories. It seems that the insula’s role may be more general: Liberati et al. suggest that it is

A hub connecting sensory areas to other networks involved in the processing and integration of external and internal information… [and may form part of] a multimodal salience network [to] allow gaining a coherent representation of different salient conditions, including, but not limited to, pain-related experiences.

The authors do note, however, that they did not record signals from individual neurons. Rather, the local field potentials that they measured reflect the firing of all of the whole population of cells close by the electrode. So it remains possible in theory that pain-specific neurons exist in the insula, mixed up with other ones.

What about other “pain-specific” cortical regions? Perhaps they will need a rethink too. While the nerve pathways by which pain signals arrive in the brain via the spinal cord are well understood, and while pain-related subcortical brain regions have been established, the question of whether any cortical areas are pain-specific seems more difficult to answer.

Just a few weeks ago for instance, neuroscientists Lieberman and Eisenberger claimed, on the basis of fMRI evidence, that “The dorsal anterior cingulate cortex is selective for pain”. However this paper was roundly criticized by Tal Yarkoni, the neuroscientist who developed the software used to perform the analysis. Yarkoni argued that other categories of stimuli activate that area, not just pain.

It seems to me that if the pain specificity of the insula and the dorsal cingulate are in question, we may need to rethink the whole concept of the cortical “pain matrix”. This “matrix” is an idea which can be found in many neuroscience textbooks:


Now, no-one is denying that these regions do respond to painful stimuli. The question is whether they are, in any sense, functionally specific to pain. If not, it might be more appropriate to refer to the matrix by another name – perhaps (although this would have problems of its own) the “salience matrix”?

ResearchBlogging.orgLiberati G, Klöcker A, Safronova MM, Ferrão Santos S, Ribeiro Vaz JG, Raftopoulos C, & Mouraux A (2016). Nociceptive Local Field Potentials Recorded from the Human Insula Are Not Specific for Nociception. PLoS Biology, 14 (1) PMID: 26734726

CATEGORIZED UNDER: fMRI, papers, select, Top Posts, Uncategorized
  • Pingback: 1 – The Myth of the Brain’s Pain Matrix?()

  • non_sig

    Some people want to name the “pain-matrix” into “neuro-matrix” :)

    I think, this means for fMRI-studies on pain-perception that they (in most/a lot of cases) should have salient stimuli and stimuli of the same kind as the pain-stimuli but in a non-painful intensity as a control. They often don’t have this, probably (often) because it would be difficult to see any activation (in the pain-contrast) then. But it must be difficult to find these stimuli (because what is the same salience?), and maybe it would be an endless task, because pain has probably more aspects then salience and intensity…

    • Paul Mullins

      Salience and intensity are often linked.

      • non_sig

        Yes, I was just thinking that it might be useful for (some) fMRI-studies to compare the pain-stimulus with a non-pain-stimulus of the same …kind (don’t know the correct term for that, sorry). Like a warmth stimulus in comparison for a (painful) heath stimulus. (For some pain-stimuli this is probably not easily possible, because they are not just some other simuli at higher intensity.) Of course the salience of the pain stimuli will be higher as well, so the intensity would not be the only difference. Therefore another comparison with another simulus of a different kind (like they did in that study) that has about the same salience (as the pain-stimulus) might be useful.
        But of course it depends on the question of the study if it needs thouse controls or not.

        • Paul Mullins

          Most fMRI studies of pain do have the “non-noxious” control stimuli. Some also have other conditions to account for the act of rating the pain, the salience, attention etc. Part of the original concept of the pain Neuro-matrix was that the physical properties of the painful stimuli alone where not enough to explain the experience of pain, and that other subjective factors also have a role to play.

          • OWilson

            Which can lead to certain conclusions about pain as pleasure, and pleasure as pain.

            Gets a little complicated!

  • Pingback: The Myth of the Brain's Pain Matrix?()

  • Greig Starton

    If you need good mmj strains and other meds contact 6013404352 or email

    .com and you will get respond asap..

  • Paul Mullins

    The “pain neuromatrix” as originally described was not an actual physical network of regions, but more a psychological concept to illustrate that our perception of pain varies based on a number of other factors like mood, energy state, attention, social situation etc. It has since been co-opted by Neuroimagers as a short hand for those regions of the brain involved in pain perception – which unsurprisingly are also involved in general perception. It’s one of the first “myths” I “bust” with any of my students who are working on pain. Like the concept that we only use 10% of our brain it is a popular fiction. Unsurprisingly most pain researchers are aware of this “fiction” and are probably aware that while certain brain regions reliably activate for pain, forming a “network” often the only thing that changes perception to pain is intensity of stimulus.

  • CL

    Thanks for bringing this up, one would hope that science has moved form finding specific regions involved in specific functions to a more networked approach several years ago. I believe most pain researchers have made this transition, but a few stragglers remain.
    I doubt that there is any area that is specific to an actual function at all. Much like Jan Smedslunds argument that psychology “cannot be an empirical science”, we are stuck with language and disagreements of language when we try to assign a brain region to a specific “function” (like the silliness of equating pain with social exclusion, or suggesting that there are no distinct emotions, because they all engage the same brain regions). A simple visual task appears to activate most of the brain (
    That said, of course there are regions that are specialized.

    • Maia

      Several upvotes for the several crucial points made in a single comment.

  • Ivan Dimkovic

    Could it be that the “sore” aspect of pain is almost completely “handled” by the subcortical structures (e.g. sensory maps in PAG, RAS, etc.) and, as such, does not require any “dedicated” cortical machinery compared to e.g. touch or sensory aspect of pain, which require more advanced analysis and cortical transformations but share the common cortical circuits due to the similarities?

    I think pain is an ancient (evolutionary) feeling which was probably one of the first aspects of whatever primitive consciousness our ancestors evolved. Could it be that most of the processing for the “sore” aspect of pain evolved before evolutionary “explosion” of neocortex, while the more advanced processing which required neocortical involvement simply was too similar (in both origin and information content) and evolved as parts of the other cortical “hubs” such as S1, dACC, Insula etc.?

    In this case, what is processed by the cortex during a painful episode would be things like e.g. location discrimination, “higher” emotional processing (fear, recalling previous similar events) and executive planning, and all of these would be “sharing” parts of the cortical mantle with other non painful stimuli.

    I am not a neuroscientist, and above idea is just a wild hypothesis based on “armchair” thinking, so it could be dead wrong.

    But I wonder if it is possible to actually “simulate” a painful stimuli by cortical stimulation only (e.g. by TMS) – not just the memory of pain and response “as if” it happened, but the actual sensory aspect of “soreness”, for the lack of better word (a pain ‘qualia’ if you will)?

    It is certainly possible to do so for e.g. vision or hearing, where TMS of V1 cortex results in really “seeing” phosphenes, indicating that the important part of visual “quale” is actually originating in activity of V1 neurons. Can the same be said for pain?

  • Celia

    I think that this makes sense and I do agree with you that the pain matrix is probably a myth. However, I have just read this paper “An fMRI-based neurologic signature of physical pain” published in 2013 in which the authors claimed to have identified a signature of pain based on fMRI data and that they are able to distinguish with this signature nociceptive from non-nociceptive stimuli. This goes against Mouraux’s studies saying that both stimuli could activate the pain matrix and that it is just about saliency. So how to explain both and get them coincide… Thanks in advance for your help.

    • Giulia Liberati

      Dear Celia, thank you for your comment. In the study you are referring to, indeed, Wager and collaborators find that thermonociceptive stimuli are consistently associated with activation in several brain regions (including the insula, S2, the thalamus, and the ACC). However, as the activation of these brain regions is often found in response to innocuous stimuli as well, we must be careful not to rely our conclusions mainly on a reverse inference. The likelihood of the inference being correct depends on the degree of exclusivity of the relationship between the cognitive state and activation in the brain region. In other words, it is true that nociceptive / painful stimuli consistently elicit activation in those regions, but we cannot conclude the reverse, i.e., that if we see activation in those region, then, the subject must be feeling pain.

      The recent paper by Hu & Iannetti (2015), “Painful Issues in Pain prediction” published in Trends in Neurosciences gives a good explanation of the interpretative difficulties of relying on “pain signatures”.

  • Giulia Liberati

    Thanks again Neuroskeptic for your thoughtful post. I am happy that it has triggered some interest!

    Here are a few papers that tackle the transition from the concept of “neuromatrix” to the one of “pain matrix”, as well as the “saliency vs. pain” issue:

    “From the neuromatrix to the pain matrix (and back)”
    Iannetti & Mouraux 2010, Exp Brain Res

    “The pain matrix reloaded: a salience detection system for the body” Legrain et al. 2011, Prog Neurobiol

    “A multisensory investigation of the functional significance of the “pain matrix””
    Mouraux et al. 2011, Neuroimage

    It’s true that most researchers are aware of the pain matrix “myth”. However, the expression is still used relatively often in the literature (as well as in conference presentation), oversimplifying the issue of how pain/nociception are processed in the human brain.

  • Pingback: The one about my first pain paper –

  • Pingback: New Study Casts More Doubt on Notion of the Brain's 'Pain Matrix' | wisdomparliament()

  • Pingback: New Study Casts More Doubt on Notion of the Brain’s ‘Pain Matrix’ | Don Santana()

  • Pingback: New Study Casts More Doubt on Notion of the Brain's 'Pain Matrix'()

  • Pingback: New Study Casts More Doubt on Notion of the Brain's 'Pain Matrix' - Gizmodo - Self Help Education Arena()

  • Pingback: Scientists Have Been Using a Flawed Method to Diagnose Pain  | JogleApp()

  • Pingback: Scientists Have Been Using a Flawed Method to Diagnose Pain  – TechINFUSE()

  • Pingback: Scientists Have Been Using a Flawed Method to Diagnose Pain  | Anish E J's Blog()

  • Pingback: Scientists Have Been Using a Flawed Method to Diagnose Pain  – Brain Sock()

  • Pingback: Scientists Have Been Using a Flawed Method to Diagnose Pain  | Big Box of Geek()



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