Let's talk now about the eye implant you are developing. Firstly, what do you hope to achieve with it?

Well, there are numerous conditions that cause blindness. Some of the more prevalent ones involve the retina itself, killing the photoreceptors – the things that are capable of receiving light and changing that into a response in the back of our head to say, 'Oh, there's light.' But if there is any nice thing about a disease, it is that this disease kills those photoreceptors only. There are basically 10 different layers of the retina. In retinitis pigmentosa and a couple of other related diseases, the nerves that are not the photoreceptor nerves basically stay alive. And when you stimulate them with electricity, they actually do whatever they used to do when you saw light. We wanted to implant some electronics (which we eventually shrunk into a silicon chip) within or around the eye, to stimulate the nerves of the retina and give some approximated vision back to some of these patients who still have the capability of seeing but no longer have the mechanism that starts it off.

But surely that won't suit people who have always been blind – none of it works.

It depends on how they went blind. The optic nerves of people with glaucoma, for example, are essentially dead. The retina itself becomes part of the optic nerve, and so those people would only be treated by a device called a cortical stimulator: a rather risky prospect in which you actually put implants in the brain to stimulate the nerves. (People who are blind don't have an awful lot to lose in eye surgery, so the risks are relatively low with stimulating the retina cells.)

How much sight might be restored – and how can we know?

How much sight do you think you could restore? Can you give some examples of what people might be able to see?

I think that a patient fitted with this is going to have to relearn how to see, just as cochlear implant patients have to relearn how to hear. And it's not going to be the way you and I see; it's not going to be 20/20 vision. Think of being at a sporting event where you can certainly see the numbers on the scoreboard, and where on some of the larger scoreboards that show pictures as well – rudimentary images, animations, that sort of thing – I guess you could make out faces if you had enough training. We may be able to get people to read; we may be able to get people to recognise at least that there is an object there, a person perhaps, movement, that sort of thing. But we are a long way off from being able to convey real images. Even so, it would be wonderful to be able to get past an obstacle without ramming into it. That is a big plus.

Light and dark perception is something that we are almost certain we are going to be able to convey. A lot of blind people suffer from horrible sleeping disorders because they lose synchrony with the 24-hour day. They don't know there is sunlight there, so they don't know when to wake up or when to go to sleep, and they shift into and out of synchronisation with us. Some of our studies have shown that the device we have is capable of at least evoking a physiological response in the brain that says, 'Yes, there is light there.' And because that is in the absence of any other light source, it has to result from our device.

So far your tests have been on sheep. How do you know what they are seeing?

The animals can't tell us what they see, but we can measure the electrical activity in their brains. When we stimulate the retina, that activity travels down the optic nerve to the centre of the brain, where it splits off and is processed in a number of ways before going to the vision centres of the brain, the visual cortex. If we can find where the visual cortex is on an animal (it's fairly well defined) we can put some recording electrodes immediately above that portion of the brain and when we stimulate it, some event is going to result if the animal is realising that it 'sees' something. And when we deliver numerous repetitive stimulations, the animal will come up with some sort of electrical pattern in its brain to tell you, ah! it sees something.

But to know what people are actually seeing – how big the light is, how bright it is, that sort of thing, which is called psychophysics – we are going to have to get some humans to tell us.

How can an eye implant become a beneficial reality – and when?

You say the device itself is a silicon chip. How big is it?

The chip can be extremely small. We are constrained more by cost than anything else, because it was an inexpensive process we happened to make a fairly large chip, about 6 millimetres in size. Beforehand, we built a similar circuit about half a metre by half a metre that had all of the components – you can go down to Dick Smith Electronics and buy the components there. We were able to shrink that to something substantially smaller, which we could put in or around the eye region in such a way that you probably wouldn't be able to tell by looking at a person fitted with this thing that they have it.

What obstacles have to be overcome before this can be implanted in a human eye?

There are some technical things. Engineering-wise, we have to make sure that this device is sealed. If any piece of electronics – a radio, say – gets into salt water or even around salt water, eventually it will stop working. The body is made mostly of salt water, so putting electronics in the body is a very difficult problem. And the signals that this chip is able to deliver have to be able to get out from the little implanted capsule to the electrodes that stimulate the nerves. That is a particularly difficult prospect, but we are making good progress on it.

There are also some medical aspects. We have serious problems with being able to fit substantial things within the eye. I have heard the retina likened to wet tissue paper. It is extremely delicate. So we have to develop techniques of surgery. We have made some good progress, actually implanting a few devices within the eye of an animal, and it seems to be a reasonable prospect to go into longer-term tests.

Probably one of the trickiest prospects is to convey the world image in a very pixelised version. By way of analogy, although cochlear implants have been around for 10 or 15 years, there is still a huge amount of activity going on, to figure out how you get sound that goes into a microphone into something that can actually stimulate people to hear. A similar thing is going to be happening with vision.

Using the retina itself we are, in fact, at an advantage, because it is all mapped – what you see on one side is actually on the other side of your retina. So we know where we stimulate and where we see. But there is still a long way to go to stimulate what we see: does it get bigger, does it get smaller, does it hurt, that sort of thing.

What is holding up the testing in humans that you need for that?

Patients overseas have been anaesthetised on the operating table and tested. But we felt that we could not yet say, 'If we get these results, then we can actually put a implant in someone's eye.' Other studies have been extremely preliminary, and we have been reluctant to use humans for things that would not benefit anybody. We want to be ready to start implanting things for real, not just for testing.

Focus questions

• What do you think Suaning meant when he said that people would need to ‘relearn how to see' once they were fitted with visual implants?

• What are the similarities and differences between the visual implant being developed by Suaning and the cochlear implant used to treat deafness?