Published on June 20th, 2013 | by Jo1
The Bionic Eye
By Andrew Cookson, Monash University
Imagine how the future will look. You may imagine phones getting even smaller and more powerful, interactive technology making our daily lives simpler, or a new level of connectedness among people all over the world. But what if the future meant granting millions of vision-impaired people one of the basic things many of us take for granted – our sight? This is the long term goal of the technology we call the “Bionic Eye”.
As a Mathematics PhD, I have been working with an inter-disciplinary team on Monash’s own Bionic Eye project. It works like this: firstly, the vision-impaired individual wears a special set of glasses with a video camera attached, which takes a video image of the surrounding area. Then, the image is transformed into a digital signal which is carried wirelessly to a chip implanted in the visual cortex at the back of the brain. Finally, depending on the signal, electrodes attached to the chip stimulate the brain in a certain pattern, creating the sensation of vision.
A fair question to ask would be: “what is a Mathematics PhD doing working on this project?” The answer is that if there were only a handful of electrodes, and a handful of brain regions to stimulate, advanced mathematics may not be as necessary. However, we have a 25 x 25 array of electrodes (that’s 625 in total), and even more potential brain regions to stimulate. Thankfully, even if we model the electrode-brain interface by trying to minimize the current used and maximise stimulation accuracy for a given target, this is a simple optimization problem to solve. What makes it difficult is that the brain has a complex response to being electrically stimulated, in which the brain cells – or neurons – will fire off their own signals, depending on time and applied current. This phenomenon is modelled by differential equations that we can use to build a simplified model of the brain’s response to stimulation. Using this model, we can make a more accurate estimation of the electrical input required at each point – the more complex the model the more accurate for optimization.
This type of model is not limited to the Bionic Eye project in its applications. Such an optimization model could also be used to control inputs for Parkinson’s disease, Alzheimer’s disease, epilepsy, depression treatment, and many more.
It must be noted that this technology is still in its infancy, and it will be a long time before we know how close we can get to normal vision. One thing is certain, though – that many of those who once thought that vision was an impossible dream, now have something to look forward to.[subscribe2]