Today's Nature has an excellent piece about an increasing and currently not well-researched trend for fMRI brain scan 'neurofeedback' treatments, where the patient is shown a visual representation of the activity of a certain brain area in the hope of learning to control it.

In this case, the big idea is that a patient with chronic pain is shown real-time activity in their anterior cingulate cortex, an area in the frontal lobe associated with the 'unpleasantness' of pain (rather than just its physical sensation), and they can see when they doing something to successfully reduce the activity and can try and learn to do it reliably.

The article looks at the work of Sean Mackey who researches the area but is appropriately skeptical about a number of companies who have recently sprung up offering this as a treatment, despite the lack of firm evidence.

As you may recall, this premature commercialisation is a bit of a pattern with fMRI research, as you can also buy the services of companies offering 'lie detection' and 'neuromarketing' despite a similar lack of eviden