![](https://swap.stanford.edu/was/20151211033644im_/https://neuroscience.stanford.edu/sites/default/files/styles/large-scaled/public/hankgreely.png?itok=9ck8VY6w)
Hank Greely
By Hank Greely
The recent meeting at Harvard on neuroimaging, pain, and the law demonstrated powerfully that the offering of neuroimaging as evidence of pain, in court and in administrative hearings, is growing closer. The science for identifying a likely pattern of neuroimaging results strongly associated with the subjective sensation of pain keeps improving. Two companies (and here) recently were founded to provide electro-encephalography (EEG) evidence of the existence of pain. And at least one neuroscientist has been providing expert testimony that a particular neuroimaging signal detected using functional magnetic resonance imaging (fMRI) is useful evidence of the existence of pain, as discussed recently in Nature.
If nothing more is done, neuroimaging evidence of pain will be offered, accepted, rejected, relied upon, and discounted in the normal, chaotic course of the law’s evolution. A “good” result, permitting appropriate use of some valid neuroimaging evidence and rejecting inappropriate use of other such evidence, might come about. Or it might not.
We can do better than this existing non-system. And the time to start planning a better approach is now. (Read on for more on how)
As I have written before, two different issues are key to the appropriate use of new scientific evidence. First, the science has to be well demonstrated and, second, the procedures for applying that science need to be routinized and widely adopted. Neither is conceptually difficult, but nor will either spring full grown into being.
The science will require large-scale trials of promising neuroimaging techniques for detecting pain. These trials should start with a condition that seems susceptible to such detection and that is commonly contested in courts or administrative agencies. Lower back pain might be a good first target. A few thousand people – affecteds and controls, of different ages, ethnicities, genders, mental health status, drug abuse history, and other likely relevant variables – should be scanned to see how well neuroimaging can determine whether they are in pain. The accuracy should be calculated, not just in terms of the tests’ sensitivity and specificity, but, whenever possible, taking into account likely test groups, the tests’ positive and negative predictive value.
If an approach passes an initial assessment that the tests can work well enough to be useful, agreement needs to be reached about the proper procedures to use in the testing. What kinds of neuroimaging devices should be involved? How long should the test be? How should it be conducted? For what kinds of people has it been, and not been, validated? Further trials could usefully shed light on these issues, in the way that Phase 3 trials for drugs exist in part to refine the doses, timing, and other methods of using a drug preliminarily found safe and effective in Phase 2. Ultimately, facilities conducting such tests should go through an accreditation process to make sure they understand, and are applying, these consensus best practices in how to do the testing.
Do these two steps seem utopian? They are what forensic use of DNA went through in the 1980s and 1990s. The science was tested and proven, in part relying on two different National Academy of Sciences committee reports. Then the procedures were hammered out, routinized, and embedded in accreditation requirements for DNA forensic laboratories. There is no reason that, if the science is promising enough, the same progression cannot happen in neuroimaging for pain detection.
One big question is who will pay for the research – and the negotiations on appropriate methods – necessary. The Social Security Administration, which must make disability determinations based on claims of pain hundreds of thousands of times a year, might be one source. Private insurers could be another. The National Institute of Justice, which has an interest in reducing litigation by providing more evidence to lead to settlements that forestall, or cut short, litigation is a third. Even the plaintiffs’ bar might be willing to pay for a seat at the table – it is, after all, in their clients’ interests, too, to have a faster and more accurate of proving their claims.
We should be laying the foundations today for such an effort. We should be at least sketching out what kinds of trials would be needed and how much money they would likely cost. We should be discussing how uniform procedures could be agreed to, looking to precedents from DNA evidence. And we should be starting conversations with the institutions and interest groups that might be able to make this happen.
We have a real chance to get in at the beginning of the law’s use of what may be an extraordinarily important technology. The steps to be taken are not within the expertise of any of us – neuroscientists or law professors. But the opportunity to something truly valuable is great. And if we do not seize it, who will? Anyone?
I hope the Harvard meeting, along with the meeting a year earlier at the University of Maryland, will eventually be remembered as the start of a time when scientists and legal experts, working together, helped get a new form of scientific evidence right. And, in doing so, improved the world.