Monday, March 11, 2013

Neuroethics Journal Club: Imaging Pedophilia

The form of argument is familiar: X is bad, very bad. We can maybe stop X, but in order to do so, we’ll have to compromise our values just slightly. We’re not happy about it, but after all, X is very, very bad.

At this month’s Neuroethics Journal Club led by psychology graduate student Katy Renfro, X was pedophilia, and the topic of discussion was a new technique to decrease pedophilic sexual desire through fMRI-based brain computer interface (BCI). The title of the paper –Renaud et. al’s “Real-time functional magnetic imaging—brain–computer interface and virtual reality: promising tools for the treatment of pedophilia” – immediately raised two questions in my mind (which, as I eventually realized, were probably not the right ones). First, is it going to be possible to Clockwork Orange pedophiles? Second, if so, should it be done?

A treatment for pedophilia?

So What Did They Do?
Nothing, actually. As one participant put it: “this is basically the beginning of a fishing expedition.” This particular fishing expedition was for a new, fMRI-integrated version of an old technique – biofeedback – in which a patient receives physiological information in order to help control subtle thoughts or behaviors. For instance, a person suffering from anxiety might practice relaxation techniques while having their heart rate and blood pressure measured. This physiological feedback can help indicate which techniques are most successful, helping the patient to better manage their anxiety.

In the context of pedophilia, Renaud et. al suggest a biofeedback technique premised on activation of the anterior cingulate cortex (ACC). The ACC, the authors argue, is associated with impulse control and sexual arousal. If pedophiles could learn to decrease activation in the ACC, then, it follows that sexual attraction to children (or at least the impulse to act on that attraction) might be diminished.

So how to decrease ACC activation? The authors propose placing pedophiles in an fMRI (voluntarily, of course), showing them computer-generated images of children (the “virtual reality” from the paper’s title), and mapping the children’s movements onto ACC activation. So as ACC activation decreases, a computerized child might raise its hand, and the pedophile would know that they had successfully decreased ACC activity. To help things along a bit, the authors also suggest a “strategy of covert mental rehearsal,” meaning the patient continually thinks about the terrible things that would happen to them if they were caught acting on their desires.

The theory behind biofeedback.
Does It Work?
I should reiterate that Renaud et. al didn’t actually do any of this: they just proposed the methodology. On the fMRI biofeedback fishing expedition, they haven’t even gotten around to buying the fishing poles. But there’s enough detail in their methodology that it’s possible to guess at the likelihood that it would succeed, something that the discussion participants were happy to do.

Concerns fell into a few categories. First, there was widespread agreement that ACC activation is too simplistic to be a reliable measure of sexual desire. Not only is the neurobiology of sexual desire poorly understood in general, but the ACC is involved in many non-sexual functions including error detection, empathy, and suppressing undesirable thoughts, functions that may not be desirable targets of biofeedback. Second, attempts to modify sexual desire, most notably homosexual desire, have historically been unsuccessful. One participant noted the difficulty this poses for fMRI biofeedback: either pedophilic sexual desire is basically similar to other forms of human sexual desire, in which case any form of “conversation therapy” should be expected to fail; or, pedophilic sexual desire is fundamentally different from other sexual desire, in which case there is no reason to target regions like the ACC that have been implicated primarily in studies with non-pedophiles. Finally, participants raised concerns regarding ecological validity. A man in an fMRI scanner, looking at computerized images of children whose movements he partially controls with his brain, wearing a cylinder around his penis to measure arousal and a camera tracking his eye movements, in a room with psychiatrists and neuroscientists and fMRI technicians, is not the same as a real-life encounter with a child. There is reasonable concern, then, that fMRI biofeedback for pedophilia would not be effective in the real world even if data showed it to be effective in the laboratory.

Perhaps not the same as viewing an emu in the Australian wilderness.
But What If It Did Work?
Well, it still might. It’s probably unfair to criticize a person’s fishing technique before they’ve even gotten to the lake, and Renaud et. al’s paper is a proof-of-concept proposal, not a meticulously worked out solution. Several participants suggested that fMRI biofeedback is worth pursuing, even if it only starts paying material dividends five or ten years down the line. So it’s worth asking: if such a technique could be used successfully with pedophiles, is there any reason why we should be concerned about implementing it?

fMRI biofeedback is not Clockwork Orange: nobody’s eyes are forcibly pried open, nobody is socialized through the infliction of horrific pain. The sense in which it’s like Clockwork Orange, and in which it might make some people uncomfortable, is that it uses technology to change behavior. There’s nothing new about this – all drugs, all therapies set out to change behavior – but there are plenty of reasons why super-effective behavioral modification techniques are scary. Luckily, fMRI biofeedback doesn’t seem to meet that criterion, so there’s probably little risk of slipping into a behavioral-control dystopia.

A more relevant concern may be the implications of successful fMRI biofeedback on how the law understands and punishes sex offenses against children. In a previous Neuroethics Blog post, Cyd Cipolla has argued that the perceived special heinousness of sex crimes against children, along with pedophilia’s incurability by any current therapy, has resulted in a set of laws premised on the notion that pedophiles “are not on the extreme edge of criminal behavior we imagine them to be, but are, in fact, in a category of their own.” These laws include provisions for lifetime surveillance and community notification following certain types of sex offenses, as well as the possibility of indefinite civil commitment for some offenders. According to Cipolla, the real issue at stake is not criminal responsibility – which is a given – but empathy: “How does knowing about the cause… help us to move towards real empathetic solutions even while maintaining guilt?”

There are still reasons for caution, however. To paraphrase one participant: we need to reflect on whether our goal is really create effective therapies, or if we’re ultimately more interested in “making ourselves feel better.” Therapy for pedophiles is enormously attractive in that it allows us to tell ourselves we’re doing something about a behavior that, for most of us, is horrifying. But to feel like we’re doing something about pedophilia is not the same as actually dealing with it – and especially in politics, the two can easily become conflated. The barriers to fMRI biofeedback are significant: we don’t know that much about sexual desire, we don’t know that much about pedophilic sexual desire in particular, there are plenty of questions about fMRI reliability, and fMRI feedback in particular is a totally new and untested technique. There are no clear answers, but plenty of questions; and given the sensitivity of this particular issue, there is a danger that limited empirical results could be blown up into significant political pressure to implement fMRI biofeedback on a large scale. That would be a shame, if for no other reason than it would discourage us from asking difficult questions about how we can most effectively deal with sexual violence.

No comments: