« Finally! Why AlGo Took So Long to Leave | Main | Unconscious Discrimination Twenty Years Later »
Tuesday, August 28, 2007
Pain Control Using fMRI
A couple of days ago, the NYT ran an article on the use of real-time functional magnetic resonance imaging (fMRI) to treat pain and perhaps a host of other symptoms like addiction and depression. The technology works like a kind of high-tech biofeedback:
Here’s how Omneuron uses fMRI to treat chronic pain: A patient slides into the coffin-like scanner and watches a computer-generated flame projected on the screen of virtual-reality goggles; the flame’s intensity reflects the neural activity of regions of the brain involved in the perception of pain. Using a variety of mental techniques — for instance, imagining that a painful area is being flooded with soothing chemicals — most people can, with a little concentration, make the flame wax or wane. As the flame wanes, the patient feels better. Superficially similar to an older technology, electroencephalogram biofeedback, which measures electrical feedback across multiple areas of the brain, fMRI feedback measures the blood flow in precise areas of the brain.
By giving users feedback about their pain, the technique attempts to create a visual representation of an individual's pain. That's pretty impressive! But imagine if we could make interpersonal judgments of pain. That could really change the way we identify malingerers and the way we calculate damages in court. As I've noted, I think that neurotechnologies may someday move us in that direction. (X-posted.)
Posted by Adam Kolber on August 28, 2007 at 11:00 AM | Permalink
TrackBack
TrackBack URL for this entry:
https://www.typepad.com/services/trackback/6a00d8341c6a7953ef00e54ee789b08834
Listed below are links to weblogs that reference Pain Control Using fMRI:
Comments
Daniel,
Of course not: Done.
Posted by: Patrick S. O'Donnell | Aug 29, 2007 2:34:03 PM
Hey Adam,
I absolutely agree that fMRI measures correlates. I just think that if we have sufficiently reliable correlates, that's all we need!
Need for what? And what is the criteria for sufficient reliability?
If I look at an X-ray and see a large fracture in a person's leg, I can make a pretty good inference that the person X-rayed is in pain.
I'm less certain of the quality of the inference. It's not unheard of for even large fractures not to result in much pain. In addition, because people's pain responses are so different, what would be agony for one person might simply be a dull ache for another. This demonstrates the roughness of the correlation, and, tracking Patrick's suggestion, the excessive caution that is needed in inductively moving from the identification of a neural correlate to a conclusion that the subject of the neuroimaging is experiencing the phenomena asserted.
My own personal opinion is that both scientists and the media generally do not exercise the caution warranted. It becomes all too easy to elide the notion of correlation with ontology, as Illes et al point out.
Does the X-ray measure pain? That's an interesting question. Does a thermometer measure temperature or just a correlation between heat and the properties of mercury?
I think there's an easy answer to the first question: no. That does not mean it fails to provide indicia of pain, but the measurement of pain, if it exists via X-ray, is an exceedingly rough heuristic.
I appreciate this inquiry but I don't need to settle it in order to make the claim that fMRI (or the next neurotechnology) can allow us to make more objective inferences about people's pain than we can now.
I respect your position, but I'm less sure that it doesn't implicate some key premises in your argument. What, for example, would it mean to draw some "more objective inferences about people's pain"? And how exactly would neuroimaging permit this? My point here is that fMRI does not do anything more than enable us to garner information about neural correlates of pain. This, of course, is no mean feat, but interpretation of the meaning of the correlation is still required both on an aggregate and on an individual level, and there is a significant danger, IMO, in presuming that blood oxygenation levels tell us a great deal about a person's pain qualia.
Bear in mind, I'm not remotely opposed to the use of fMRI to improve diagnosis and even treatment of pain. But I do think it is crucially important to identify what we are actually measuring (and what we are not measuring), in large part because I am going to argue that it is primarily our cultural attitudes towards subjectivity that lie at the heart of our consistent problems in adequately treating pain. Thus, thinking that objective evidence of pain will permit better treatment, while not necessarily false, raises a number of questions regarding why we need what we deem to be objective evidence of pain in order to treat it adequately.
Patrick,
Would you mind terribly emailing me the Bennett and Hacker citation? Thanks!
Posted by: Daniel Goldberg | Aug 29, 2007 1:52:45 PM
Hilary Putnam has offered arguments against even the notion (the picture) of psychophysical correlates (as is typically understood) in his latest writings and the assumption that "psychological states must be pared down to an 'internal core' before they can be metaphysically and scientifically [and legally!] legitimized" is likewise arguable and a topic Putnam has also addressed.
As Bennett and Hacker explain: "The observed neural phenomena that are concomitants of a person's suffering pain...are not forms of *pain-behaviour.* [Keeping in mind that the 'primary grounds or evidence for the ascription of psychological predicates to another are behavioural.'] They are *inductively* correlated with being in pain. The correlation is an empirical discovery, which presupposes the concept of pain and its nexus with criterial, non-inductive evidence for the application of the concept of pain to *a living creature* (not to its brain)." We cannot ascribe psychological attributes to the brain, and I think it is mistaken to believe that "the 'access' which introspection gives each person to his own mental states and processes *is* a limited access to processes going on in his brain," such that the fMRI is somehow going to give us a truer, more objective picture of the experience of pain, one whereby we can detect, for instance, "malingerers." I rather agree, again, with Bennett and Hacker: "The sense in which there is literally any such thing as 'getting inside another human being' is: examining the interior of his body and brain--for there is no such thing (save figuratively) as getting inside his mind. This is not something we cannot do; rather, there is no such thing to do." Getting inside another person's brain, literally or through images of one sort or another, does not tell us what a person is thinking or feeling. "Moreover, given that we can and do establish that different people enjoy or suffer the same experience, differences in the underlying neural processes would not show that, despite their sincere avowals to the contrary, they do not really have the same experience. To think otherwise would be to suppose that someone might sincerely avow pain, groan and scream, but, nevertheless, that a brain scan migh show that the person was not in pain at all. But that is absurd. All that would thus be shown is that a neuroscientific theory, which assumes neural uniformity across different individuals, is false. For the criterial (logically good) evidence for being in pain, which is required for the non-inductive identification of the pain of others, is presupposed by inductive evidence correlating pain with neural states and events, and accordingly overrides it."
Sunny Auyang has also reminded us how the experience of pain is not merely a raw sensation but something close to an emotion" "Besides the ouch the pain has an *affective aspect* that determines its unpleasantness: how badly it hurts and disrupts other mental processes." Indeed, this helps to account for the fact that "mental concentration can sometimes block pain."
The criteria for identifying pain are not located in the head.
Posted by: Patrick S. O'Donnell | Aug 29, 2007 1:23:42 PM
Daniel: I absolutely agree that fMRI measures correlates. I just think that if we have sufficiently reliable correlates, that's all we need! If I look at an X-ray and see a large fracture in a person's leg, I can make a pretty good inference that the person X-rayed is in pain. Does the X-ray measure pain? That's an interesting question. Does a thermometer measure temperature or just a correlation between heat and the properties of mercury? I appreciate this inquiry but I don't need to settle it in order to make the claim that fMRI (or the next neurotechnology) can allow us to make more objective inferences about people's pain than we can now.
Posted by: Adam Kolber | Aug 29, 2007 11:42:46 AM
As I've mentioned to Adam over email, I'm skeptical of the claims that fMRI provides objective evidence of pain. As far as I'm aware, fMRI measures correlates, which is not equivalent with the qualia under investigation. Evidence of pain is certainly not the same as the experience of pain itself, especially because evidence may be of varying quality.
Judy Illes et al. refer to this erroneous conflation as a "neurofallacy," and specifically the neurofallacy of "neurorealism." I don't necessarily agree with Adam that the neurotechnologies may move us in that direction, if only because I agree with Searle that there is an ineluctably subjective feature to consciousness (and a fortiori pain) that, IMO, by definition could not be captured by objective instrumentation.
One question which I am answering in my dissertation is why we feel the need to seek objective evidence of pain.
Posted by: Daniel Goldberg | Aug 29, 2007 9:32:58 AM
Thanks for your comment Jim! There are a number of ways to manipulate fMRI readings, so I don't know whether it will ever be practical to use it to detect malingering. In particular, most litigation is likely to concern pain that occurs over a long time, while fMRI is likely to focus on pain at a particular moment.
I do reference a study in the paper that looks at structural features of the brain to assess chronic back pain. It would be much more difficult to fake a measurement of a structural feature of the brain. But you are certainly right that courts already try to make inferences about people's pain, and it's a very inexact effort. As for pain tolerance, I've thought about that too. I'm afraid I don't have much of an answer yet--it cuts to the very heart of what it means to be in pain.
Posted by: Adam Kolber | Aug 28, 2007 4:02:58 PM
If one can make the flame wane with a little concentration I would imagine that a malingerer could just as easily make it wax. This sounds like self hypnosis made easy. Doesn't George Clooney use a similar technique since his injury during the making of Syrianna?
Seriously though, objectively measuring pain with technology I think raises some interesting questions (which you may address in your paper, the paper I am about to read). For example, would the technologies be able to take into consideration one's ability to tolerate pain? And even more interestingly to me, should one's ability to tolerate pain matter? Plaintiff's attorneys can and do point to the plaintiff's outward expressions of grief as justification for a larger pain and suffering award. The implication is that that particular plaintiff is actually suffering more than one who might tolerate the loss better.
Posted by: Jim Green | Aug 28, 2007 3:23:10 PM
The comments to this entry are closed.