RECOGNITION that genetic tests were open to abuse in the 1990s led to a frank public debate about the ethics of testing and who should have access to the results. And while not all the problems have been resolved, at least we now know what the dangers are. Another area that is crying out for a similar debate is brain scanning.
Researchers have gone beyond looking at how a generic human brain works and are eagerly trying to identify variations in the workings of different people鈥檚 brains. This week, New 杏吧原创 carries three articles that describe the rapid pace at which the frontiers of brain scanning are moving (see 鈥淏ehind the mask鈥�). Researchers are delving into the most intimate details of who we are, including such things as our personality traits, moral reasoning and tendency to violence. While their aims are commendable, many of their results raise big questions. If the scan of an individual with a tendency to violence looks 鈥渁bnormal鈥�, what does that mean? How reliable is the finding, and who should be told?
New 杏吧原创 has argued before that brain scans purporting to show a link with racial prejudice must be taken with a pinch of salt (22 November 2003, p 3). It is still unclear if these scans show inherent racism in action or its conscious suppression. This highlights one of the big problems with brain scans 鈥� although they show basic processing in the brain, it is not at all clear what they tell us about a conscious, thinking human.
Advertisement
There are other reasons to be cautious about scans. They present findings in a seductive way that can exaggerate the certainty of those results. They come, after all, from sophisticated multimillion-dollar machines, so how can they be wrong? If only things were that simple. A functional MRI scanner is not a simple spring balance, where what you see is what you get. It does not provide you with a direct view of the brain, as some proponents of scanning argue. Images of the brain emerge only after a huge amount of processing has been done on the raw data, and the nature of that processing is often chosen by researchers to present results in a way they understand.
There are potential dangers here. Take the thresholds for each coloured region on a scan, which have to be chosen. Other branches of science have shown that the value chosen for thresholds can fundamentally change the nature of results. Epidemiologists, for example, have found that when looking at small geographical areas of, say, 5 kilometres in diameter, a cluster of leukaemia cases can look highly suspicious. Yet pull out to 10 kilometres and the cluster can disappear 鈥� the number of cases in this circle appears perfectly normal. What would small changes in colour thresholds do to the appearance of an 鈥渁ctive鈥� amygdala or 鈥渓azy鈥� ventral putamen?
Even so, scans are undoubtedly alluring for the uninitiated 鈥� another reason to be wary. Data about someone鈥檚 motivations or desires could be of real value to those trying to detect criminal intent or detain people who may pose a threat to others. But it is hard to come by. The temptation to use brain scans where no other data exists is likely to increase, even though the information contained within scans is often imperfect.
Likewise, faced with the option of a scan or the results of a psychological test, it is worth betting that most people would give more credence to the scan. The irony here is that psychological tests are often used to make sense of fMRI data 鈥� to decide how raw data is processed.
Already marketing moguls and political spin doctors are starting to use brain scans to try to sharpen up their advertising. How soon before lawyers start to see scans as useful additions to courtroom evidence? Before that happens, neuroscientists must be certain of what their scans can and cannot reveal, and just as important, they need to convince the public of their case.