Can we really train our brains? The hope that maybe we can is based on the idea of “brain plasticity”: that no matter how old you are, and how set in your ways, your brain is constantly changing to adapt to new experiences.

Brain plasticity is based on a decade’s worth of hard science, which is how it came to spawn a billion-dollar brain training industry. But there’s a problem. When scientists look at whether brain training games actually make people smarter, the answer is very clear: no.

I have spent the past two years trying to find out why this is and what, if anything, we can do to get the best out of our brains. Armed with a shopping list of improvements, from extending my attention span to upgrading my abysmal sense of direction, I went looking for science-based improvements. After being zapped with magnetic pulses, strapped into wearable brain aids and taught how to meditate, I’ve concluded that while brain games might make you better at the task you are practising, they won’t improve your overall brain “fitness”.

But a new wave of neuroscience might offer a better way. In recent years, neuroscientists have been exploring the brain’s connectome – its wiring diagram – with the aim of finding out how areas are linked into networks, and how activity in these networks differs from person to person. This is important because differences in ability are likely to come down to how well the brain can combine information from different areas to create something that is greater than the sum of its parts.

With this in mind, rather than aiming to strengthen particular areas, it makes more sense to work on learning to engage different circuits of the brain depending on the job at hand. This might sound harder to achieve than a quick go on a brain training game but, in my experience, it can be done.

Working with scientists at Boston Attention and Learning Lab, I found that learning to focus for longer was not, as I imagined, about building strength in the circuits needed to concentrate. It was about learning to allow my brain to fluctuate between focusing hard and wandering off, which uses two different circuits (the dorsal attention network and the default mode network respectively).

The training, which involved clicking on some images and ignoring others for 12 boring minutes at a time, was about practising balancing these two mental states to stay just focused enough. After trying it, I found the same skill applied to swimming in a pool on my own and in some forms of meditation. It could work for anything that takes focus, but not too much: perhaps playing an instrument, making something with your hands or singing in a choir.

Similarly, I found that improving my maths ability wasn’t about doing numerical press-ups to get better at sums, but learning to control my terror of doing them. My fear was diverting brainpower from the circuits I needed to solve the problem, so the scientists in Boston zapped part of my brain’s emotional-control circuitry with electricity to dial it down, allowing me to divert power to where it was needed. Zapping isn’t strictly necessary – they told me slowing down and relaxing would probably do much the same.

With this kind of research into how to engage our brains more effectively coming through, it feels premature to throw the idea of brain training out completely. Meaningful change is possible, but it will come, not from tinkering with the engines of our minds, but by learning how to drive.

Caroline Williams is the author of Override: My Quest To Go Beyond Brain Training and Take Control of My Mind (Scribe, £14.99). Order a copy for £12.74 at bookshop.theguardian.com