noigroup logo

Bold phantom claims

By Timothy Cocks Science and the world 24 Jun 2014

Interesting article that relates to a paper that was published online in Pain.

Solved: Mystery of phantom limb pain

“In a way, the debate about the origin of phantom limb syndrome is a version of the ancient mind-body problem,” said professor Marshall Devor of the Hebrew University of Jerusalem’s Department of Cell and Developmental Biology and Center for Research on Pain, who led the study. “Our research resolves the problem, at least in this case. It’s the body.”

The ghost in the machine

Skeptical of conventional brain-centric explanations of phantom limb syndrome, Devor and his colleagues set out to test another neurobiological theory: that phantom limb syndrome comes from the nerve fibers that used to run to the amputated limb. Guided by medical imaging, the researchers injected 31 leg amputees who suffered from phantom limb syndrome – 16 in Albania and 15 in Israel – with local anesthetic near where the nerves from their amputated legs enter the spinal cord in the lower back.

Within minutes, phantom limb sensation and pain was temporarily reduced or eliminated in all the amputees. Control injections had no effect; nor did numbing nerves in the stump in the few cases tested.

Body over mind

“The dismantling of phantoms via the silencing of these ‘terror cells’ is in my mind a death blow to the brain theory of phantom pain and a call to industry to join forces with scientists to find new treatments for neuropathic pain,” said Dr. Haim-Moshe Adahan, who ran the Israeli trial at Chaim Sheba Medical Center at Tel Hashomer’s Pain Rehabilitation Unit and is working on a follow-up study to extend symptom relief using steroids.”

These are bold claims indeed. The final quote above from one of the researchers is certainly very strong with it’s reference to the “death blow to the brain theory of phantom pain”.  But the ‘terror cells’ (what a terrible phrase to use) can’t produce pain – they can only send erroneous nociceptive signals to the brain, but what about the idea that “nociception is neither sufficient nor necessary for pain”? Do these results suggest sufficiency of nociception for pain?

The actual paper published in Pain was not quite so strident in its discussion of the findings.

“Our data are inconsistent with maladaptive cortical plasticity being the primary driver of PLP and npPLS. However, some of the secondary peculiarities of phantom limb sensation (eg, telescoping and reference) may well reflect plasticity of cortical processing. Had the impulses interpreted by a conscious brain as PLP originated in the cortex, spinal and intraforaminal blocks would have been ineffective and certainly not topographically appropriate. Beyond that, the very foundations upon which the cortical plasticity hypotheses rest are equivocal.

We propose that ectopic PNS discharge, primarily that originating in DRG serving the amputated limb, drives CNS somatic representations to generate a conscious percept of the phantom limb. The quality of the sensation, PLP or npPLS, presumably depends largely on the types of primary afferent neurons that contribute to the ectopic barrage”

There are a number of very strong dualisms at play here – mind/body, brain/body, peripheral nervous system/central nervous system, top down/bottom up. The whole dualistic argument just evaporates though if you consider one continuous nervous system without arbitrary slash marks separating peripheral from central and one whole human being with an embodied mind embedded in an environment and culture. Then we can say with certainty that bodies don’t have phantom limb pain, brains don’t have phantom limb pain and minds don’t have phantom limb pain – only people experience phantom limb pain.


-Tim Cocks


Get up to date and get you think on at a noigroup course or immerse yourself in some brainy books with Explain Pain 2nd Edition or The Graded Motor Imagery Handbook.


  1. Beautifully put Tim that only beings have pain not body parts……….Maybe the researches are threatened by having to take the responsibility of the pain experience upon oneself rather than giving over the responsibility to God…….Religion verses science, the old dualistic battle still raging in the 21st century……….it reminds me of the Spanish Inquisition as depicted by Monty Python……
    DB London

  2. Gday Tim,

    I’m pleased that NOI will openly look at opposing views. It’s a REALLY important study and I’m stunned that this hasn’t been done before… or has it?

    I can’t get the full text, but I’d like to know:

    1) how significant were the changes in pain?
    2) was this a properly blinded trial (double or triple)?

    If the researchers giving the injections were not blinded, then their expectations could very easily create a reduction or obliteration in experienced pain in subjects. That would make it a brain output, not a peripheral mechanism.

    The ‘pain as output’ theory is still very solid in my way of thinking UNLESS this study was triple blinded… in which case the theory could be under threat. The blogs on hypno-analgesia and placebo are very strongly suggestive that pain is an output.


    1. Hi EG
      1) The reduction in PLP was very significant – in one group (n=11) spinal (intrathecal) blocks using lidocaine resulted in 100% of the participants reporting at least 75% attenuation of their PLP and 90% of participants reporting at least 75% attenuation of their non-painful phantom limb sensations. Intraforaminal blocks of 13 participants led to 90% reporting that they had experienced at least 75% attenuation of their PLP.

      2) The participants in the trial were blinded to the substances injected at times, as well as when the actual injections were given at others. The protocol for the injections seemed a bit variable, and the authors do state that “we acknowledge that our observations fall short of accepted criteria for randomised blinded placebo controlled trials. For practical reasons, we did not include an experimental arm randomized to placebo injections
      exclusively, and although patients were blinded to the type and timing of the injections, the medical staff was not.” However they also state “Spinal and intraforaminal block consistently attenuated, and often completely eliminated, both PLP and npPLS in lower-limb amputees. Control injections did not…. Sham injections and intraforaminal injections of nonblocking solutions never blocked PLP, even when patient were intentionally told to anticipate a pain-relieving injection. Blocking solutions, even when
      injected covertly, consistently did.”

      I do think this is important research and the findings are worth pondering and discussing. Reading the article, my understanding is that the authors main point is that “maladaptive cortical plasticity” or “multi sensory mismatch” are not the primary driver of PLP, and rather that “ectopic PNS discharge, primarily that originating in DRG serving the amputated limb, drives CNS somatic representations to generate a conscious percept of the phantom limb. The quality of the sensation, PLP or npPLS, presumably depends largely on the types of primary afferent neurons that contribute
      to the ectopic barrage”. So not so much that the brain isn’t involved in PLP, but that the primary driver is the peripheral ectopic discharge.

      So even here pain, it would seem, can still be considered as an ‘output’ or construct, yet despite this, the authors comments, at least in the news article, were incredibly dualistic and very strongly confident in declaring an “answer” to the so called mind/body problem in the case of PLP.

      Interestingly, Phantom limbs have been reported in 20% of a population of people with congenitally deficient limbs with 20% of this group reporting that they experienced PLP (Melzack R, Isreal R, Lacroix R and , Schultz G.
      Phantom limbs in people with congenital limb deficiency or amputation in early childhood Brain, 120 (1997), pp. 1603–1620). And PLP treatments targeting representations of the limb such as GMI have been shown to have efficacy – do these people with PLP have no DRG ectopic barrage? Or does the ectopic barrage stop?

      In the end, i still reckon the dualisms inherent in the paper and news article are ultimately unhelpful and more likely to distract from potential collaboration and work towards better answers for people experiencing pain.


  3. So what?
    Let’s see what happens to those patients phantom limb pain after a few months.
    Every peripheral nerve block type of surgery is effective-for a short period of time.
    The pain usually returns after a few weeks or months.
    And what about brachial plexus injuries?
    The too have a kind of PLP.
    This is just another one of those “we have finally solved the problem” studies that you never ever see or hear about again.
    Brains are more complicated that one would like them to be.😊

    1. Hi Matthias
      Thanks for stopping by and commenting.

      Some of the participants had catheters inserted with infusion pumps and the authors stated that these participants reported that the attenuation of PLP was maintained for up to 12 days, but, there was no long term follow up and i strongly suspect that the PLP would have returned.

      There was a sentence in the paper that left me quite saddened – After completion of the procedure “They were then released. Because of the inaccessibility of many of the subjects, the study focused on the short-term effects of the blocks. We did not attempt systematic long-term follow-up, although sporadic feedback was obtained from some subjects.” I’m assuming that ‘inaccessibility’ means that the participants may have been from rural or remote areas, perhaps with limited access to support. So here is a group of people, most of whom have had limbs blown off by land mines, who for a fleeting period of time received some relief from their pain and were then ‘released’. No follow up, nothing to take home and continue that might help to relieve pain and probably no hope. I wonder if many of them fully understood what it was that they had just gone through?

      Brains and people are certainly more complicated than we would like to think!


  4. adambjerre

    Great pragmatic debunking, Tim. Some dualisms in science are pretty secure (matter/anti-matter, living/non-living and hardware/software) while the ones you mention are misleading and ties pain on the wrong level. (Consciousness of) pain is indeed at network property that cannot be reduced.

    Andy Clark might suggest that the researches are essentially fiddling with prediction error while thinking that they are actually “turning the knob (down)” on pain.

    Keeping up with the conceptual hygiene: what do you mean when you say “people *have* phantom limb pain”?

    Thanks for your great work on the blog.

    1. Thanks so much Adam
      I reckon at least the living/non-living duality may still be a little shaky – viruses (What little I understand of them – seem to inhabit a bit of a grey area between these two poles!

      “Conceptual hygiene” – I love it, and you are right, people can’t “have” (possess) a “thing” called pain – slippery thing language and metaphor, so easy to slide into non-trivial category mistakes.

      I’ve edited my post, replacing “have” with “experience” – this is the least wrong way that i can think of – do you think there is anything better?


      1. adambjerre

        Thanks for the article, Tim. At least according to Dennett viruses are not alive. They are strings of nucleic acid with attitude. 🙂 But they are as apt for natural selection as any living thing. That is pretty much what I know about viruses.

        “Pain is slippery.” – L. Gifford

        I myself fall into slippery language all the time. Pain does not fall easily into the ontology of “the scientific image” (Sellars). On the one hand we cannot talk about pain as a thing (like molecules and atoms) and on the other hand we don’t want to say that there is no such thing (eliminativism).

        In relation to pain how do we make a more internalist sub-personal account of sensory transduction go together with the conception of personal-level experience involving real-world action? As far as I can see we have two kinds of candidates of explanation to take seriously: 1) the enactivist conception put forward by Varela, Thompson, O’Regan and Noë among others and 2) a generative model of perception put forward by Friston, Hohwy, Roepstorff and Clark. My guess is that the enactivist perspective would regard pain as an experience, while the generative model would regard pain as nociceptive prediction (that can go wrong), but that might be a huge simplification on my behalf and I apologize.

  5. Thanks for doing that detailed summary Tim (and for sending the original to read).

    Patients were blinded, as expected. The timing of injection was also blinded which is very good practice. But the fact that the researcher wasn’t blinded introduces a potential for confusion. A researcher with a certain look, certain way of breathing, certain posture, certain vibe will tip the patient off that something powerful is happening or about to happen. Next thing you know, the pain has disappeared as a placebo response.

    The other possibility, as you mention is: “not so much that the brain isn’t involved in PLP, but that the primary driver is the peripheral ectopic discharge”.

    Maybe there’s a 3rd possibility too.



  6. Those injections do pose a risk – for transforaminal steroid injections this applies, (might be different for lidocaine?)

    Cervical Transforaminal Epidural Steroid Injections: More Dangerous Than We Think?

    Conclusions. This study demonstrates a significant risk of serious neurologic injury after cervical TF-ESIs. A growing body of evidence supports an embolic mechanism, whereby inadvertent intra-arterial injection of particulate corticosteroid causes a distal infarct. Embolism to the distal basilar artery region can cause midbrain, pons, cerebellum, thalamus, temporal and occipital lobe infarctions. Other potential mechanisms of infarction include vertebral artery perforation causing dissection/thrombosis and needle-induced vasospasm.


    With those potential risks I’d pass.

Your email address will not be published. Required fields are marked *


Product was added to cart.