Expecting bad things – what are the repercussions?

I am currently on the train to Wauchope, NSW to visit my husband who is doing a rural medical placement. Now in my head, I decided that train food would be shocking and so when low and behold, I got my meal, I was pleasantly surprised to discover that it was not only edible but quite…delicious?  I know! It knocked me for one too. However, this is something that I’ve noticed – if I don’t have high expectations of things, I am often pleasantly surprised and it makes for a happy day. However, if I expect everything to be amazing then I’m constantly disappointed (like when your friends tell you it will be the best movie ever…).

So how surprised was I to discover that expectancy seems to play the opposite role in terms of pain – eg, if you expect/believe/are conditioned to expect a stimulus to be less painful, it usually is less painful. A recent narrative review in Neuroscience Letters from Atlas and Wager has amalgamated the literature (mainly experimental) on the role of expectancy in pain.[1] This fab review considered placebo responses (eg, expectations that a treatment will cause pain relief, even though the treatment is inert, results in pain relief), nocebo responses (eg, increase in pain that accompanies beliefs that a treatment will cause pain or increase symptoms), and stimulus expectancy effects (eg, expectations about noxious stimuli that can be elicited through instructions or arise from spontaneous inferences or can develop though associative learning processes – see Daniel’s post on classical conditioning!). What this review does is give a great overview of the various responses, body systems, and brain and spinal cord regions possibly involved or related to expectancy. Now what I plan to do is just cherry pick a couple of points that I hope will raise some discussion.

My first point relates to the expectancy effects in experimental pain versus clinical pain. First, it is well-established that the placebo effect occurs in clinical conditions (see classic review here[2]).  However, much of the brain imaging data that provides the neurobiological link between pain reports and actual changes to the nociceptive circuitry comes from experimental pain studies (note: the review[1] does present some interesting data on irritable bowel syndrome[3] and fibromyalgia[4]). Certainly there are many practical reasons for this, but the fact remains that we would likely expect somewhat different responses in people with chronic pain than with experimental pain. Therefore while these studies definitely give us a much better understanding of the role of expectancy in experimental stimuli, we are lacking crucial research into clinical pain. I posit that this becomes particularly important when we consider the difference in emotional impact between pain that is transient vs pain that you don’t understand and that doesn’t go away.

My second point relates to how we develop theories regarding expectancy and pain – specifically how we base a lot of theories on correlations. We often assume the relationship between 2 variables goes one way or another (ie, inappropriately inferring causation). First, assuming causation is not the way to go, but regardless, we tend to speculate the direction of a possible relationship because it helps us design future studies. So what if relationship direction that we choose is wrong? For example, we find that people who tend to have the best responses to placebo have higher levels of trait optimism.[5] Thus there is a tendency to think, if I just become more optimistic, I will get a better response to placebo (kindly ignore the complete oversimplification of optimism being the only factor related to a robust placebo effect). However, what if the relationship was actually in the other direction? What if all my life I just tended to have good responses to everything (eg, my brain worked differently and processed information differently) and so I became quite optimistic, because in fact, for me, any treatments for pain always ‘worked’. Then we have mistakenly given optimism a key role in placebo response vs inherent differences in brain processing of noxious stimuli being the key. Please note that the review did not suggest this but it was just something that made me think.

What would be fascinating to me would be to run a longitudinal study where we image people and then follow them for ages (ie, until enough develop chronic pain) and then re-image everyone. At baseline we could image cortical responses to a noxious stimulus and cortical response when we induce positive expectations (ie, pain relief) for experimental pain. Then when we can image the same things once a person has developed chronic pain which would allow us to compare their ‘new responses’ to their ‘old responses’. We could also image cortical activations during their chronic pain and also activations in response to a placebo treatment aimed at their chronic pain, allowing us to evaluate differences between chronic pain and experimental pain (in one person). I think a longitudinal study would allow us to create a better overall picture. Anyone have some spare pocket change?

So based on this review paper[1], I come away with a healthy respect for the role of expectation in pain. However, regardless of the results of this review, I think I’ll stick with having no prior expectations going into movies…the alternative is just too disappointing!

Tasha Stanton

Tasha Stanton Body In MindTasha Stanton is a postdoctoral research fellow working with the Body in Mind Research Group both in Adelaide (at University of South Australia) and in Sydney (at Neuroscience Research Australia). Tash has done a bit of hopping around in her career, from studying physio in her undergrad, to spinal biomechanics in her Master’s, to clinical epidemiology in her PhD, and now to clinical neuroscience in her postdoc. Amazingly, there has been a common thread through all this hopping and that common thread is pain. What is pain? Why do we have it? And why doesn’t it go away?  Tasha got herself one of the very competitive Canadian IHR post-doctoral fellowships and is establishing her own line of very interesting investigations.  Her research interests lie in understanding the neuroscience behind pain and its clinical implications. She also really likes nifty experiments that may have no clinical value yet, but whose coolness factor tops the charts. Last, Tash is a bit mad about running, enjoying a good red with friends and organizing theme parties. Tasha, aka Stanton Deliver, was the all round best performer at the Inaugural BiM Table Tennis Comp. Here is Tasha talking more about what she does and a link to her published research.


[1] Atlas LY, Wager TD (2012). How expectations shape pain Neuroscience Letters, 520, 140-48 DOI: 10.1016/j.neulet.2012.03.039

[2] Beecher HK. (1955). The powerful placebo. Journal of the American Medical Association, 159, 1602-606 DOI: 10.1001/jama.1955.02960340022006

[3] Price DD, Craggs J, Verne GN, Perlstein WM, & Robinson ME (2007). Placebo analgesia is accompanied by large reductions in pain-related brain activity in irritable bowel syndrome patients. Pain, 127 (1-2), 63-72 PMID: 16963184

[4] Goffaux P, de Souza JB, Potvin S, & Marchand S (2009). Pain relief through expectation supersedes descending inhibitory deficits in fibromyalgia patients. Pain, 145 (1-2), 18-23 PMID: 19524367

[5] Morton DL, Watson S, El-Deredy W, Jones AKP. Reproducibility of placebo analgesia: effect of dispositional optimism. Pain 1999; 80: 1-13.


  1. Tasha Stanton says

    Steve and Ali,
    I reckon we all need to have a good discussion over a vino in Milan!! Thanks guys for your great comments.
    Ali – it’s funny that you bring up Bingel 2011 – that remifentanil study is currently my favorite study! Fascinating! However, it would be interesting to have them run it again and randomise the order of the different conditions (ie, positive, negative, no expectation) to control for order effects. Thanks for the great links!

  2. Ali Khatibi says

    Steve and Tasha..this is a really encouraging discussion between you two and I am enjoying reading your thoughts on this issue. Regarding the brain imaging help in this field, it seems that we are looking for cognitive modulation (Or as Stev mentioned the term Bias) of mechanisms involved in nociceptive and analgesic processes, and we would like to have an objective measure of this effect using imaging techniques. The challenge arises when some researchers argue that we normally over-discuss findings of imaging studies and generalize those findings to a level in which it is not applicable at that level (as Tasha mentioned in her latest comment)(See: Logothetis (2008) : What we can do and what we cannot do with fMRI).
    On the contrary, we have nicely designed studies investigated cognitive modulation of analgesic process using imaging studies. In this area many work have been by Irene tracy and Katja Wiech at FMRIB and Bingel at UKE, ahmburg for example:
    Bingel et al., 2011: The effect of treatment expectation on drug efficacy: imaging the analgesic benefit of the opioid remifentanil.
    Bingel et al., 2012: Neuroimaging as a tool to investigate how cognitive factors influence analgesic drug outcomes.
    Tracy 2012: Getting the pain you expect: mechanisms of placebo, nocebo and reappraisal effects in humans.
    and this is link to one of Irene’s presentations on this topic titled:
    Can Imaging Dissect Analgesic versus Placebo Responses?

    Lauren Atlas Reply:

    Thanks for sharing the Tracey slides! See also our recent paper using the balanced placebo design and remifentanil: Though analgesics and placebos both reduce pain, they actually do so through separate brain pathways, with additive effects:

    Now I’ll stop plugging our work, but both were super relevant to your discussion, I hope!

  3. Tasha Stanton says

    I take your point Steve and I vaguely remember that presentation. However, if in all conditions people are in pain and rating it, changes in imaging findings should reflect whatever is different between the 2 conditions, such as differences in expectations (or unreliable measures!). I certainly don’t think imaging tells us the whole story (far from it) but I reckon it becomes an interesting supplement to behavioural or perceptual measures.

    Steve Kamper Reply:

    But how do we know which part of the pretty colours on the image are due to pain and which parts are due to thinking about pain? Potentially people with different expectations could have very different tendencies in this regard. Of course, as you point out this also ignores any issues with reliability of the measures.
    This feels like a conversation that we can take up over a Peroni and/or a Vino Rosso in Milan.

  4. Steve Kamper says

    I wonder if we can untangle the reporting bias issue using imaging though. My recollection is far from perfect, but I seem to remember a presentation (at Aus Pain Society Conf. in Gold Coast I think – James can you help me out?) where the authors showed similar imaging findings in people with pain, as those who were asked to think about scoring pain on a rating scale.
    Anyway, it certainly seems like a technology that is moving forward quickly, so maybe this space will look very different in a couple of year’s time.

  5. Tasha Stanton says

    Hey Steve,

    Thanks so much for your comments – insightful as always! Thanks for the reference to the newer placebo study (sometimes I just can’t resist citing oldies). I think you really bring up a good point regarding terminology and definitions (ie, placebo effect – effect without a mechanism). Probably because I only dabble in this area, rather than being a proper placebo researcher (haha!), I would often combine the idea of expectation and placebo, although I certainly do think there are ways to separate them. I guess I never got hung up on the idea of an effect without a mechanism (ie, if you stick to the definition of placebo religiously) just because intuitively that doesn’t make sense to me (ie, of course there has to be something causing that effect). You have given me some food for thought on this one.
    Regarding the brain imaging, of course we have ages to go! But I was suggesting less that we image the ‘placebo response’ as you suggest, but rather image nociceptive responses, which we now know quite a bit about. If we can image nociceptive responses to noxious stimulus or maybe even fluctutating clinical pain levels and then we change expectations of that person (by instructions or suggestions for example) then we can image changes in the nociceptive circuitry. This I think can help guide us in some ways. But I do agree with you in that it is not the be-all end-all technology. To me it just helps rule out the possibility of reporting bias as I feel it is much stronger if pain scores change and they are accompanied by changes in brain activation.
    Great comments as always Steve-o and sorry I’ve been tardy in replying!

  6. Tasha Stanton says

    Cool! Sounds good Ali! I’ll definitely come check your poster out and I look forward to meeting you!

  7. Steve Kamper says

    Nice work Tash, thanks for posting, I think this really is a fascinating area. I have a couple of observations:
    In line with your musings in point 1, I would tender Hrobjartsson and colleagues’ SR (New Engl J Med 2001 and subsequent updates) on placebo effect sizes in clinical trials as a replacement for Beecher’s tired old piece from 1955. They found that placebo effect sizes in clinical trials are really small (maybe 5 or so on a 100 point pain scale). So the question of what is going on re: placebo effects in trials (and maybe in practice?) isn’t quite the open-and-shut case we used to think it was. Wrt brain imaging results (while agreeing with you regarding the debatable clinical relevance of most of these studies), I’ll stick my head out here and ask the question whether we know enough about the tool, how it works and what it means to learn too much from it about placebo effects just yet?
    I’m glad you bring in the theoretical issues in point 2. It’s a massive problem in the field. Most of the quantitative work published regarding patient expectations in clinical research is more or less atheoretical, or at least doesn’t stand on a theory that is expressed explicitly. It consists mostly of secondary analyses of datasets where the researcher happened to have stuck an expectation measure in somewhere. Not that this can’t tell us something, but it does leave a bit of a mess if you’re trying to sort through the literature and work out what the hell is going on!
    Last point before I shut-up (sorry, but I told you I like this stuff). I wonder if the idea of the placebo effect itself is reaching its use-by date. Here you seem to be talking about expectations and placebos more or less synonymously, some people would agree with that, probably others wouldn’t. I think part of the problem is that the placebo effect as we generally think of it is something of a contradiction in terms. A placebo is something that is inert, yet a placebo effect really happens – it is an effect without a mechanism. I would favour letting the magical placebo black box slip into the realms of medical history. If we think there is an effect, then there must be a reason for it, let’s use the scientific arsenal we have at our disposal to find out what it is so we can use it for the benefit of patients. Of course, the problem with this is that I would no longer be able describe myself as a placebo researcher, and I’m always reluctant to destroy a lame academic joke.

  8. Tasha Stanton says

    Fantastic comment – thanks for contributing! Many of the issues that you raise (ie, attential bias, uncertainty tolerance) were also raised as key areas for future research by the review authors. It becomes such a complex process to tease out the different effects and make sure that our findings are telling us what we actually think they are telling us! Could we include a conditioning effect on one body part (ie, induce expectations of threat) but then move on and induce expectations of threat on another body part (no conditioning – just a one-off thing), but apply the noxious stimulus to the first body part (ie, not expecting to receive it here and attention is actually directed to the other body part)?

    Ali Reply:

    Thanks for your encouraging feedback. In my idea the design you mentioned is a very nice way to evaluate the effect of expectation and to control the effect of attention direction, but there are some point which in my idea my raise some concerns:
    1- when we condition subjects at the first level, implicitly we manipulate their interpretation, and this could influence their understanding of a painful stimulus (see Keogh and Cochrane, 2002: Anxiety sensitivity, cognitive biases, and the experience of pain).
    2-the focusing and distracting itself might influence processing of painful stimuli (Nouwen et al., 2006: Effects of Focusing and Distraction on Cold Pressor–Induced Pain in Chronic Back Pain Patients and Control Subjects)
    3-to control the effect of focusing and distracting we can manipulate attention using a parallel task which trough that we can control the exogenous and endogenous orienting (see : Van damme and Largin, (2012): How efficient is the orienting of spatial attention to pain? An experimental investigation.)

    Tasha Stanton Reply:

    Ali – are you going to be at the IASP meeting in Milan? Would be great to chat! Thanks for the refs – will take a look!

    Ali Khatibi Reply:

    My poster number is PT 474 (Tuesday-morning session)… I already have your poster in my itinerary, that would be great if we can meet there, maybe we can brainstorm on this topic and reach to some nice design. I am really looking forward to meeting you in Milan.

  9. Sara Brentnall says

    Nice work Tash!

  10. Annie Tucker says

    Thank you! As a chronic, neuropathic pain patient and researcher-in-training [graduate student]; I so appreciate your articles with references!

  11. The expectations are linked to the anxiety and how much uncertainty people can tolerate. In research on the role of expectancy in cute pain conditions many investigations was done in dental pain (These authors have the most reputation: Arntz, van den Hout, Rachman). Many of these studies confirmed the mismatch hypothesis which itself can challenge the findings on the placebo effect. The mismatch hypothesis claims that under-predicted pain will be followed by an increase in predicted pain while the placebo effect predicts the reverse pattern. On the other hand, the role of cognitive processes (like attentional bias) in these processes has been neglected. There are studies showing that expectation of threat increases attentional bias toward source of threat, but there is no systematic investigation on the interaction between placebo effect, attentional bias, and intolerance of uncertainty. These might be further investigated in future studies…I hope…

    Lauren Atlas Reply:

    Thanks for the kind post and thorough discussion of our review! Regarding the role of attention in expectancy effects on pain, we actually recently published a study directly examining the role of attentional focus in expectancy effects on pain:

    We found that indeed, threat increases attentional bias toward noxious stimuli, but that this redirection of attention actually REDUCES the effect of the threat cue itself. So the idea is that there’s actually some natural change in attention that’s protective in response to expectations of high pain, and that there’s an asymmetry between expectations for threat versus safety (high vs low pain). We didn’t look at intolerance of uncertainty, and I think that these effects might be distinct from the role of attention in placebo analgesia (sustained processes might rely on very different mechanisms from transient expectancies), but hopefully this adds a useful bit to the puzzle.

    Tasha Stanton Reply:

    Thanks for the comment Lauren!! And especially for the link to your other paper. Fascinating findings re: threat!! Perhaps see you as IASP as well?

    Lauren Atlas Reply:

    Sadly I will be missing it! I just started a postdoc in a lab that doesn’t study pain per se. It’s a shame, because the placebo session looks great and I’ve never been to IASP. Tor’s going though, and will present some new data we have been working on. Hope you guys can check it out!