8 February 2018
Andrew Masterson
Andrew Masterson is a former editor of Cosmos.
Stereoscopic vision in praying mantises works in a significantly different way to that found in mammals such as humans, and could point the way to better visual processing for robots.
That’s the finding of a team of researchers at the Institute of Neuroscience at Newcastle University in the UK – as well as the premise for perhaps the most visually appealing experiment of all time.
Stereoscopic vision allows an organism to accurately judge distances. Two eyes, facing forward but situated a distance apart, present two slightly different views.
The information from each is processed and combined in the brain, resulting in a three-dimensional image. This allows a human – for instance – to accurately judge the position of a moving ball in order to catch it, or (more ancestrally) to pinpoint the position of a river fish in order to spear it.
Stereoscopic vision in vertebrates appears to have evolved at least twice – once in mammals and again in birds. Although it is comparatively rare in birds – think owls – it is present in most mammal species, except whales and dolphins. In some, such as humans, other primates and cats, it is highly developed.
One theory holds that stereoscopic vision evolved among mammals because our early ancestors were nocturnal, and 3D perception combined with high visual sensitivity were necessary adaptations for surviving in the dark.
For whatever reason, however, it also developed in praying mantises. The team at Newcastle University, led by behavioural ecologist Vivek Nityananda, decided to find out if it worked in the same way as it does in birds and mammals.
They did this by training a mantis to wear a pair of rose-coloured glasses and watch movies.
The test mantis, resplendent in its new spectacles, was shown film of a prey species. The glasses made it appear as if the two-dimensional image was in fact three-dimensional and hovering right in front of the insect.
The mantis was evidently fooled by the illusion and tried to catch it. Having established that their system worked, Nityananda and his colleagues then screened mini-movies comprising dot patterns that had previously been validated in research investigating human stereoscopic vision.
They discovered that mantis sight works very differently to our own. Humans, for instance, can be shown two slightly different still images – as in the nineteenth century amusement, the stereoscope – and interpret the result as a single 3D picture.
Mantises, it turns out, can’t do that – probably because it’s an ability that brings no fitness advantage for them. Instead, mantis vision depends on identifying which bits of a visual field are moving. The bits that aren’t are of no interest.
“This is a completely new form of 3D vision as it is based on change over time instead of static images,” says Nityananda.
“In mantises it is probably designed to answer the question ‘is there prey at the right distance for me to catch?'”
For robotics developers, this may turn out to be very good news indeed. Current attempts to develop stereoscopic vision in mobile robots are all modelled on the system used by the human brain. As such, they are complex, clumsy and very energy-hungry.
Modelling a system on mantis vision – where the unmoving sections of the visual field are irrelevant and only the much smaller areas that aren’t static require processing – is likely to be faster, and cheaper.
“Since insect brains are so tiny, their form of stereo vision can’t require much computer processing,” says co-author Ghaith Tarawneh. “This means it could find useful applications in low-power autonomous robots.”
The research is published in the journal Current Biology.