Haemodynamically responsive

Everything about functional magnetic resonance imaging (fMRI) is fascinating, except the experience of being in the narrow little tube for an hour. For the uninitiated, fMRI is a technique for observing the operation of a living brain. It does so by measuring the rate of blood flow and degree of oxygenation in various parts of the brain. This is often done while the subject is asked to perform various kinds of action, the idea being to learn more about how the brain processes information and makes decisions. For a number of reasons (not least because it is using the indirect blood measures rather than looking at actual neural activity) it is not a perfect tool for understanding brain function. It has, however, generated some really interesting findings.

While inside the machine, I had a few general scans during which I didn’t need to do anything. A couple sounded comically similar to the kind of blaring piano notes in bad horror movies. Most sounded more like buzzing or throbbing. I also had twenty minutes of playing a gambling game much like the one I played earlier outside the fMRI machine. Unlike the previous test, this one had three clear types of problems:

1. Even breaks: where you would compare, for instance, a 50% chance of winning or losing 20 points with a 50% chance of winning or losing 10.
2. Good odds: for example, ‘an 80% chance of winning 60 points with a 20% chance or losing 20′ versus ‘a 50% chance of winning 10 points with a 50% chance of losing 10.’
3. Sucker bets: for example, ‘a 20% chance of winning 20 points with an 80% chance of losing 60′ versus ‘a 50% chance of winning 10 points with a 50% chance of losing 10.’

In each of three rounds, you started with 100 points. I ended up with more about 350 in each case. Whichever of the 41 people scanned does best will win a £50 voucher. That wrinkle made me consider playing a high-risk strategy. With a low-risk strategy, one would expect the distribution of outcomes to be a curve with a fairly low standard deviation. A high risk strategy would have a similar mean, with a higher SD. If you only get a prize for being the best in a large group, it might make sense to play for the chance of winning big, even if you run a bigger risk of losing completely. Despite knowing this, I chose the safer option of playing safe on break-even situations, gambling on good odds, and minimizing risk when presented with sucker bets.

For the longest part of the scanning session, I was playing a much less interesting game. All you did was push a button and wait to see whether a slot machine would pay out or not. In this part, you got to keep whatever you won, but had no control over how much that was. After fifty minutes or so of lying motionless on my back in a noisy little tube (limbs a bit sore from fellwalking and cycling up to the John Radcliffe Hospital), I wasn’t hugely in the mood to watch random chance unfold.

After the scanning, I had a discussion with the woman running the experiment and the man who was running the fMRI machine. I was hoping to get a picture or data file including some of the data, but apparently the university has a strict policy against doing so. Too many people were self-diagnosing weird brain problems and demanding consultations with neurologists, apparently. Personally, I would have used a 3D data file to make some kind of interacting flash animation to put on my website somewhere.

At some future date, I need to go back for the diffusion tensor imaging scan. I thought all the scanning was today, but apparently neglected that there were actually two sessions.