Humans can see a single photon at a time

Humans are capable of super-sight, it seems, with a new study showing subjects in a dark room could detect a single photon hitting their eye better than chance.

Jonathan Tinsley at Vienna’s Research Institute of Molecular Pathology and colleagues in Austria and the US built a gadget able to send one photon at a time to a person’s eye or nothing at all.

When they tested three healthy men, aged in their 20s, the subjects correctly identified the presence of a photon with a probability above chance.

The experiment and device were published in Nature Communications.

Where cells in your retina called cones are responsible for detecting colour, in dim light your rods really come to the fore. There are more of them, too – some 120 million compared to six million or so cones. 

And while cones are focused in the fovea – a densely packed patch of cells in the centre of your retina – rods are everywhere else. This is why, at night, you’re able to see faint stars out of the corner of your eye, but they seem to disappear when you look directly at them.

So just how good are your rods?

Studies in the 1940s found people could detect as few as five or seven photons at a time. But the technology to test fewer photons has been lacking. So Tinsley and his colleagues built the gear – a quantum optical device that could generate a pair of photons.

One photon, called the signal, was sent to a subject’s eye, while the other, the idler, was directed to a super-sensitive detector. They were a bluish colour, around 500 nanometres in wavelength, because rods fire strongest when they’re hit with such photons.

Three men – one of whom was wearing contact lenses – were chosen to sit through a total of 30,767 trials. They sat in a dark, sound- and light-insulated room, wearing headphones. Their head was kept still on a bite bar and headrest, and were asked to fix their right eye on a very dull red light.

During each trial, the men were shown two light stimuli – one where a single photon was fired into eye into a rod-dense part of the retina and the other which was completely dark. They were asked to say which of the two had the photon.

After they gave their answer, they were told through the headphones if they were right or wrong, and the next trial followed.

Each subject sat through up to 20 sessions. But this, the researchers admit, was not enough to get statistically significant results for the individuals so their scores were pooled and analysed.

Overall, they found the probability for a correct photon detection was 0.516, or 51.6% of the time. This may seem only just above chance, but over the 30,000-odd trials, the researchers claim this is a statistically significant result.

Perhaps more interesting was the priming effect the study found. If a subject was told they successfully detected a photon, they were more likely to give a correct response in the next trials – 56%, in fact – for up to 10 seconds.