Essentially the findings show the humble honey bee can see both the forest and the trees, a skill which not even primates can boast.
Published in the journal Proceedings of the Royal Society B, the results have implications for artificial intelligence.
"This is a complete change in our thinking of how brains process visual information," said associate professor Adrian Dyer from RMIT University, a co-author of the paper.
With under 1 million neurons, a bee's brain has less neurons than a human retina. But Professor Dyer said this made it the perfect model for studying information processing at its most basic level.
"In the 1970s we saw (Star Wars robot) C-3P0 walking around and interacting with very complex things," Professor Dyer said. "But the reality is that if we think of where robots have come in the last 20, 30 or 40 years it is remarkably disappointing."
What has been holding things back is working out how to enable robots to process the complex visual information in their environment.
"Machine vision is capable of seeing the local information, or the trees, but it can't stitch it together properly to see the global information, or the forest, to know how to interact with the environment," Professor Dyer said.
Conducted over four years, the honeybee study provides great insights into the minimal neural requirements for global processing. It also proves high-level cortical processing isn't essential, meaning that the bee-brain could be a model replicated artificially.
To study the way the honeybee processes visual cues in their environment, researchers from RMIT, Monash University and the University of Toulouse watched as individual bees were released in a Y-shaped maze.
After travelling up the stem and reaching the intersection, the bees were faced with a choice of pattern at the end of each corridor. Each pattern contained both global information (from a distance the pattern formed a square or triangle) and local information (shapes within the pattern were circles and diamonds).
In subsequent tests, patterns were changed so that local and global information was swapped. The results showed the bees returned the patterns with the original global information.
It's not the first time the bee brain has proved more adept than its size might suggest. A decade ago, experiments showed bees could recognise faces. Bees, when rewarded, have also been shown to be able to learn tasks in a matter of hours.