Brain training improves your ability... to do brain training.
There has been a huge market for software that aims to claims to improve our cognitive ability by 'training' the brain to be more effective. Apparently Dr Something-or-other's Brain Trainy Thing (I may not have got the title quite right) is the most popular game on the Nintendo DS in the UK.
I have always found something slightly sinister about this kind of software. First there's the pseudo-medical/scientific implication of it being Dr Something-or-other's. The only products I'm used to being labelled Doctor Something's are quack medicine, sandals and bovver boots. Then there's the idea that by doing some basic maths and a few puzzles you can somehow train your brain. To be fair, going on the research recently carried out, it does train your brain - but only to do this particular application. It makes you better at playing this game. The 'training' isn't transferable to other cognitive activities.
This is one of the things that has always worried me about Susan Greenfield over and above the mess of the Royal Institution - she endorses quite an expensive computer brain training product called MindFit, but I have not seen any good evidence for it working, only unconvincing reports that weren't published in reputable journals.
I'm not saying you can't improve the effectiveness of your brain. Practising most things will get you better at that particular thing - this is totally proven. Being in good physical condition, with enough sleep, will help brain function (as anyone who has tried to work with a hangover can testify.) Also, you can use techniques, for instance, to help you generate new ideas more effectively - these don't boost cognitive ability, they just give you new starting points and routes to ideas. But I am always highly dubious of any claims that using some computer program will enhance your general ability to think. It's the brain equivalent of a crash diet - a supposed short cut to something that really needs a much broader, longhaul approach.