Change what you think, change what you feel

Cognitive behavioural therapy reduces central sensitization

Pain is not a simple sensory experience. Negative thoughts about the meaning of pain or unpleasant emotions like fear and depression can, in some cases, cause more suffering than the actual sensation. Psychological treatments like cognitive behavioural therapy (CBT) target thoughts and emotions associated with pain and have shown effectiveness in halting the cycle of depression and disability that frequently accompanies chronic pain.[1]

Despite the effectiveness of these treatments, they do not work for everyone. Improving the effectiveness of psychological therapies requires a better understanding of how they work. One key question is whether these treatments only change cognitive and emotional responses or whether they actually change the sensory experience of pain. In other words, can changing our pain-related thoughts actually change our sensitivity to peripheral input that causes pain?

Brain stem regions like the periaqueductal gray (PAG) and rostral ventromedial medulla (RVM) can send descending signals to the spinal cord to either increase or decrease the level of sensitivity to incoming nociceptive signals (the peripheral input that usually initiates pain). These regions are connected to cortical and subcortical regions involved in cognition and emotion, allowing pain sensitivity to be modulated by thoughts and feelings.[2] Consistent with this possibility, numerous neuroimaging studies have demonstrated an association between cognitive modulation of pain and coactivation and/or functional connectivity of cortical regions and descending modulatory circuitry.[3-6] What has not been established is whether this circuitry can be trained: can teaching people to alter their cognitive and emotional responses to pain actually change their sensitivity to sensory input?

To investigate this, we designed an experimental model of CBT.[7] Healthy, pain free participants received 45 short bursts of painful heat to their non-dominant forearm in each of eight sessions. We measured the intensity and unpleasantness of pain as well as secondary hyperalgesia, which is increased pain sensitivity of the undamaged area around a wound. Secondary hyperalgesia is the product of sensitization of the central nervous system. In addition to the spinal cord, this sensitization has been linked to descending modulatory regions like RVM,[8] suggesting the possibility that this sensitization could be subject to top down modulation based on thoughts and feelings. We therefore hypothesized that CBT could reduce secondary hyperalgesia.

Prior to each session, half of the subjects were given a short (five minute) session of CBT based on a published treatment manual.[9] They were encouraged to reduce negative thoughts about the thermal pain and to focus on positive aspects of the training (e.g. learning pain coping skills to be used in the future). The rest of the subjects were given training that did not focus on pain (focusing instead on improving interpersonal relationships).

As we expected, CBT reduced self-reported pain unpleasantness but not intensity, suggesting that the training altered the emotional response to pain. More intriguingly, the CBT group showed a 38% reduction in the size of the secondary hyperalgesic area while the control group’s secondary hyperalgesia was not reduced.

These results demonstrate that psychological training can reduce central facilitation of pain. This finding is clinically important because central sensitization has been demonstrated in many chronic pain conditions.[10, 11] These results suggest that reduction in such sensitization might be one mechanism whereby CBT imparts its beneficial effects. Future research might focus on how individual differences in the degree to which these responses can be trained might underlie differences in how individuals respond to CBT and other psychological treatments.

About the authors

Karen Davis

Karen DavisKaren D. Davis, PhD is a Professor in the Department of Surgery and Institute of Medical Science at the University of Toronto, and heads the Division of Brain, Imaging and Behaviour – Systems Neuroscience at the Toronto Western Research Institute.

Tim Salomons
Tim salomons

Tim V. Salomons, PhD is a Marie Curie International Incoming Fellow and Assistant Professor in the School of Psychology and Clinical Language Science at the University of Reading, UK. He is interested in the cognitive and biological mechanisms that make pain salient and how individual differences in these mechanisms might underlie differences in coping and treatment response.


[1] Williams, A. C. de C., Eccleston, C. & Morley, S. Psychological therapies for the management of chronic pain (excluding headache) in adults. Cochrane Database Syst. Rev. Online 11, CD007407 (2012).

[2]  Fields, H. L., Basbaum, B. & Heinricher, M. M. in Textbook of Pain 125–142 (Elsevier, 2006).

[3] Bingel U, Lorenz J, Schoell E, Weiller C, & Büchel C (2006). Mechanisms of placebo analgesia: rACC recruitment of a subcortical antinociceptive network. Pain, 120 (1-2), 8-15 PMID: 16364549

[4] Eippert F, Bingel U, Schoell ED, Yacubian J, Klinger R, Lorenz J, & Büchel C (2009). Activation of the opioidergic descending pain control system underlies placebo analgesia. Neuron, 63 (4), 533-43 PMID: 19709634

[5] Salomons TV, Johnstone T, Backonja MM, Shackman AJ, & Davidson RJ (2007). Individual differences in the effects of perceived controllability on pain perception: critical role of the prefrontal cortex. J Cogn Neurosci., 19 (6), 993-1003 PMID: 17536969

[6] Kucyi A, Salomons TV, & Davis KD (2013). Mind wandering away from pain dynamically engages antinociceptive and default mode brain networks. Proc. Natl. Acad. Sci U.S.A., 110 (46), 18692-7 PMID: 24167282

[7] Salomons TV, Moayedi M, Erpelding N, & Davis KD (2014). A brief cognitive-behavioural intervention for pain reduces secondary hyperalgesia. Pain, 155 (8), 1446-52 PMID: 24569149

[8] Urban MO, & Gebhart GF (1999). Supraspinal contributions to hyperalgesia. Proc. Natl. Acad. Sci. U.S.A., 96 (14), 7687-92 PMID: 10393881

[9] Thorn, B. Cognitive Therapy for Chronic Pain: A Step-by-Step Guide. (The Guilford Press, 2004).

[10] Arendt-Nielsen L, & Yarnitsky D (2009). Experimental and clinical applications of quantitative sensory testing applied to skin, muscles and viscera. J Pain, 10 (6), 556-72 PMID: 19380256

[11] Edwards RR (2005). Individual differences in endogenous pain modulation as a risk factor for chronic pain. Neurology, 65 (3), 437-43 PMID: 16087910


  1. Simon Roost Kirkegaard says

    Hi KW

    I would also be interested in hearing what John means but I expect BPS model falls down as a scientific model would be because you can quantify or compartment all the possible effects, interpretations, etc. that occurs within the BPS model which would be why most testing happens within the biomedical model and then applied within the BPS model.

    This is only a guess obviously. Maybe my interpretation is all wrong but then I would love to be corrected so I can learn more 🙂

    Kind regards,


    Simon Roost Kirkegaard Reply:

    Sorry John just saw your response with link to the article Pain Medicine and Its Models: Helping or Hindering? Reading it now 🙂

    John Quintner Reply:

    Simon, I hope we answer your question in this article. Best wishes. John

  2. Simon Roost Kirkegaard says

    Hi Tim

    I am interested in understanding how the biopsychosocial and biomedical model for pain aren’t incompatible?

    It was my believe that the biomedical model for pain or disease was outdated and no longer valid?

    Would you elaborate to help me understand your position?

    Kind regards,


    Tim Salomons Reply:

    Sure, I’m likely using the term “biomedical model” in a different way than you (and possibly a different way from anyone else, so thanks for requesting clarification). When I say biomedical and biopsychosocial models aren’t incompatible, I’m simply saying that all aspects of pain have corresponding biological processes. Pain does usually originate from nociceptive signals, but the brain can modulate processing of these signals based on the sensory, emotional, cognitive or social context.

    Where I think an emphasis on physical/biological processes goes too far is when we invalidate subjective report because we can’t directly observe (or don’t yet understand) the biological process that gives rise to a particular experience.

    Don’t know if that clarifies at all…….

    John Quintner Reply:

    Tim, when you say that “all aspects of pain have corresponding biological processes,” are you not falling into the trap of reductionist thinking? The fact that we can view the experience that we call “pain” in particular ways does not mean that the experience itself is made up of these same separable components. I seem to recall the late Patrick Wall arguing along the lines that “pain comes to us as a complete package”. But I would agree with you that a large explanatory gap remains between this particular experience and the biological processes that have lead to its creation.

    Tim Salomons Reply:

    I think it would be reductionist if we were to say that the biological level is the best level to understand a given aspect of pain, or that somehow the problem will be lest complex if we break it down to its constituent biological processes.

    I’m especially aware of this tendency as a neuroimager and a psychologist. When neuroimaging first began to explode, people had an expectation that we’d now be able to crack phenomenon whose complexity had baffled us for decades at the level of behaviour because now we were “getting right to the source”. It turns out that what’s complex at the level of behaviour is equally or more complex at the level of biology. Biology and behaviour are two different levels of analysis and which one is better largely depends on the question you’re asking.

    That said, the other extreme (that we shouldn’t expect what we observe behaviourally to have a corresponding biological process) seems dualist to me.

  3. John Quintner says

    @ KW. We did elaborate upon this theme in our 2008 paper. I have attached the link:

    Where does this lead? Hopefully it will result in a Kuhnian paradigm shift amongst pain theorists. But one never knows!

  4. Thanks Tim.

    I don’t have access. I would appreciate a link to this (also one to your website).


    Tim Salomons Reply:

    Here’s a link to my website. If you go to publications and find the paper (“A brief cognitive behavioural…..”) you’ll see the two interventions linked for download:

  5. John Quintner says

    On reading the paper, you will find that the authors make it quite clear that their study design imposes severe limitations as to how their results are to be interpreted. While their findings are of interest, I suggest that it would be premature for readers to draw any firm conclusions as to the efficacy of CBT (which to my admittedly simple way of thinking is akin to opening a “can of worms”) in comparison to other forms of treatment.

    Tim Salomons Reply:

    There are lots of studies on the efficacy of CBT for pain. In a nutshell, it works, although effects tend to be small (depending which end measure you study….pain? disability? depression?, and whether you compare to an “active control” or wait list group). I suspect that part of the reason for these small effects is that there are substantial individual differences: it works well for some people and not others. But that’s a whole other topic…..

    More importantly, I feel perhaps you’re framing it wrong by talking about its efficacy “in comparison to” other forms of treatment. I think an advantage of these types of treatments is that they can be used alongside other treatments like pharmacotherapy, physiotherapy, diet etc. In fact, an advantage of cognitive behavioural approaches is they can be used to enhance adherence to other treatments (e.g. training people to take a more positive and engaged approach to their physiotherapy or medication). Because pain is a complex experience (sensory, cognitive, emotional, social etc.), the best treatment regimens attack it from all sides…..

    Tim Salomons Reply:

    Williams, Eccleston and Morley have a good, comprehensive review on the effects of CBT for pain. PMID: 23152245

    John Quintner Reply:

    Tim, could you please tell me how anyone can “attack” a complex experience from all sides?

  6. Brilliant news on the research findings. I use CBT with my chronic pain clients and would also like to see the CBT instructions that were used in the study. Additionally, I’d be interested in knowing more about the imagery re-scripting method.

  7. This is really great research, good to hear the findings explained so succintly!
    I am researching a specific method sometimes used as part of CBT for pain management called ‘imagery rescripting’. Its basically about picturing pain, then altering that picture using your imagination.
    Would it be possible to post a link to my online research on this website or in the comments, as I really need participants?

  8. Mandy Mercuri says

    I have been using these techniques for years and they do work, thanks for proving it!

  9. Hello,

    Good study, thanks!
    Could I see the 5 minute CBT instructions please?

    Tim Salomons Reply:

    We’ve included the instructions in the supplemental materials of the paper. Let me know if you don’t have access and I’ll try to post them on my website and provide a link……

    The instructions were condensed and adapted from Beverly Thorn’s “Cognitive Therapy for Chronic Pain: A Step by Step Guide”, which is a great resource:

  10. What an excellent study, congratulations. You are so fortunate to be able to work at this fascinating interface of psychology and neurobiology. Look forward to reading more.

    Tim Salomons Reply:

    Thanks !

  11. Thank you for a interesting article 🙂

    I have a question not to the article but to the video attached to the article. I get the impression that pain originates from the body? From 1:50 there is talk about pain signals, pain information and pain circuits. There may be a nociceptive driver to a pain experience but this explanation or wording to me sounds like the biomedical and reduces the value of the great information from both the article and the video. My interpretation becomes ok pain comes from the body and can then be modulated by “other stuff”. To me it doesn’t clarify pain as a complex experience based on biopsychosocial pain model where pain is produced by the brain.

    Is my interpretation wrong?

    Thank you again for the article.

    Kind regards

    John Quintner Reply:

    Simon, I agree with your concern expressed over some of the language used in the video. It seems to me that the long discredited belief about the nervous system being “hard-wired” is proving extremely difficult to dispel.

    Simon Roost Kirkegaard Reply:

    Thank you for your reply John 🙂

    What about the authors and

    What is your position? By promoting the video it provides some validation to it and could increase the misconceptions about pain which in my practice is a huge roadblock for effective treatment. The biomedical interpretation always seem to linger.

    John Quintner Reply:

    My position is that Body in Mind is a unique forum in so far as it gives us all the opportunity to draw attention to such important conceptual errors. Hopefully the penny will eventually drop. But, as you probably know, the Pain literature contains many examples where key opinion leaders have failed to employ epistemic logic – here defined as the logical application of knowledge and reasoning to problem solving. Epistemic confusion has been the result. These problems will only be solved when journal editors and reviewers come to accept that they need to exercise a more disciplined and responsible approach when dealing with material submitted to their journals for publication.

    Tim Salomons Reply:

    To be clear, I was not involved in making the video, but I don’t see the study and the video as incompatible, nor do I see “biopsychosocial” and “biomedical” models as incompatible. There are known pathways from the periphery to the brain that are largely pain specific. There is also a great deal of interaction between these pathways and neural systems involved in emotion and cognition (not surprising since, as you point out, pain is a complex sensory, emotional and cognitive experience). I wouldn’t use the term “hard-wired” to describe pain pathways, simply because it implies that experience can’t alter transmission. In fact, probably the most interesting thing about the study for me is that it demonstrates that with time and training, changing thoughts and emotions related to pain can change sensory features……

    John Quintner Reply:

    Tim, with respect, “pain pathways” do not exist. I am not the first person to have drawn attention to this mistake.

    As I understand the situation, what is called the “biomedical” model certainly qualifies as a scientific model. But what is popularly known as the “biopsychosocial” model is really a framework or aide memoire to be negotiated and used by clinician and patient. It falls down as a scientific model.

    Tim Salomons Reply:

    yes, “nociceptive pathways” is more correct.

    KW Reply:

    “But what is popularly known as the “biopsychosocial” model is really a framework or aide memoire to be negotiated and used by clinician and patient. It falls down as a scientific model.”

    Interesting. Please explain/elaborate if possible. Where does this lead? Very interested in learning new concepts! TIA

  12. Very interesting, thanks for sharing. I have a question:

    If the CBT reduced central sensitization, then why wouldn’t this have decreased the intensity of the pain in the area of the direct noxious stimulation?


    Tim Salomons Reply:

    Thanks for the question. Primary hyperalgesia (sensitivity in the area of the wound) and secondary hyperalgesia (sensitivity in the adjacent area) are generally considered to have different mechanisms, with the former due to sensitization in the periphery and the latter due to central sensitization. So while the two might occur together, this does not need to be the case.