This futuristic-sounding scenario is hinted at in a review paper published today in the journal Neuron, which looks at how neuroimaging findings can effectively predict things like the above. The review covers over 70 studies “in which initial brain measures (neuromarkers) are correlated with or predict future education, learning, and performance in children and adults; criminality; health-related behaviors; and responses to pharmacological or behavioral treatments.”
Lead author John Gabrieli, a neuroscientist at MIT, explained that a neuromarker is “an objectively measurable indicator of a brain function or structure that is relevant for an intervention.”
He told me the idea of this kind of correlation wasn’t to achieve perfect prediction, but something that could nevertheless have meaningful impact.
For example, it’s currently difficult to tell which patients might benefit from a treatment for depression or anxiety. “The typical number you see in the literature is about 50 percent of patients have a substantial benefit from a treatment, but the other 50 percent don’t,” said Gabrieli. “That’s where we are now in a practical way.”
Similarly, it’s hard to tell which children might do well at school—and which might struggle—or to predict if a criminal is likely to be a repeat offender, or a drug addict will relapse. Imagine if there were a better way to foresee personal difficulties before any damage is done. In many cases, Gabrieli said, brain imaging offered at least a better estimate than current tools.
By looking for measures like levels of activation and connectivity in certain parts of the brain, you might be able to better decide which children might require a specialised educational plan, or make a stronger choice about which treatment to give a patient. That could save a lot of harm, and even money. As Gabrieli put it, a brain imaging session would probably work out to be less expensive than three or four extra visits to the doctor because a treatment didn’t work.
Gabrieli also pointed to a now well-known study that found a judge was more likely to grant a prisoner parole either at the beginning of a day or right after lunch. “These are the kind of factors that are influencing these decisions now, because people making these difficult decisions don’t have very good practical information about who’s more likely to offend or not.”
Of course, all of this is pretty early stages and should be considered with several disclaimers. The studies are scattered across fields and often used small sample sizes and short-term outcomes. The parole study involved eight judges, all based in Israel. In the review paper’s conclusion, the authors write that larger studies are required to transition to any sort of practical application.
And while the prospect of a neuroimaging early warning system for future failures sounds useful, there are obvious ways it could all go horribly wrong in practice. If you could tell if someone’s more likely to fail at something, what’s to stop people using that information to discriminate rather than offer help? This kind of science should not, Gabrieli argued, be used to “track and select” only those most likely to succeed.
The authors also point out the risk of diverting attention from other issues, such as behavioural and social factors; brain measures aren’t the only thing that impact an individual’s behaviour. There are also other ethical considerations to take into account, like how culpable someone is if markers in their brain reveal them to be more likely to commit an offence in the first place.
“There just has to be a concurrent discussion of the ethics—once you know somebody’s at risk, the responsibilities families and societies have to help that person,” Gabrieli said.
But he believes that, with the requisite further research, this tool could move into the everyday sphere—and soon. He pointed out that the kind of treatments and educational tools people could be guided to aren’t new, neither is brain imaging technology. “I really think it could be—man, everything is slow in progress usually—but less than ten years,” he said. “I see this as highly doable.”