Beauty is in the eye of the bee-holder

This year's IBRO World Congress of Neuroscience has a sweet treat in store - one of the plenary speakers is noted researcher Professor Mandyam Srinivasan, winner of the 2006 Prime Minister's Prize for Science.

He will explain how the study of the common honeybee is revolutionising the fields of neurobiology and robotics by simply going about their daily routine.

Vision is a complex optical and neurological phenomenon and is perfect for studying simple neural systems such as those in insects. The human brain uses the stereo optics of 'triangulation' to process information received by the left and right eye, and thereby judge distance. Insect eyes are too close together to make this system useful for their visual needs, and instead use motion cues to tell whether an object is close or distant.

Via this remarkably simple but extremely reliable, fast and accurate visual system, insects, and more specifically bees, can take off rapidly, fly through the centre of narrow gaps, navigate targets and land perfectly every time. And all of this with a minute brain, the size of a sesame seed and comprising only a million nerve cells.

The July IBRO World Congress of Neuroscience in Melbourne will host international and Australian speakers covering a staggering diversity of topics in neurobiology from synapses to psychosis. Professor Mandyam Srinivasan from the University of Queensland is a plenary speaker at the congress and will speak on Small Brains, Smart Minds: Vision, Navigation and Cognition in Honeybees and Applications to Robotics.

Srinivasan will discuss his impressive body of research on vision in honeybees that includes numerous landmark findings and itself spans basic science in engineering and biology to the latest in aeronautical robotics ... and all based around the mind of a bee.

"Bees are very easy to train, with a simple food reward, to learn new information about their environment in just a few minutes," Srinivasan says. "They are amazing." And he would certainly know - over nearly three decades, he has combined biology and engineering in thousands of elegant experiments with bees. This work has provided numerous landmark findings on visual processing in simple brains.

One of his earliest major discoveries, published in Nature in 1988, was that bees gauge distances to objects in terms of the apparent speeds of the objects' images. Paths between objects or narrow gaps are negotiated by balancing the apparent speeds of the images in the two eyes, resulting in a perfect flight path through the centre.

This 'centring response' is largely independent of object texture and image contrast, so the bees for example can safely negotiate the gap between two trees with very different barks. Srinivasan subsequently showed that bees modulate their flight speed by holding constant the average image velocity perceived by both eyes.

"This allows them to fly fast in open terrain and slow down to a safer speed in densely cluttered terrain," he says.

In the laboratory, bees flying through a narrowing tunnel slowed down as the walls narrowed and the images came closer. Several different robotics laboratories have incorporated these principles into robots that navigate through corridors for applications in the transportation and mining industries.

Coming in for landing

The way in which insects land has always particularly fascinated Srinivasan and in 2000 he described an unexpectedly simple and elegant strategy for bee landing. "The image speed of the ground or stationary surface is held constant as the bee approaches, thus automatically ensuring that flight speed is close to zero at touchdown," he says. "No explicit knowledge of flight speed or height above the ground is necessary."

In other words, as the landing point gets closer it appears to be moving faster, so the bee slows down. The result: a perfect touchdown every time - and a landing technique that Srinivasan is now trialling using model aircraft.

He also showed that foraging bees determine distance flown based on how much the image of their surroundings has moved on the retina - a sort-of visually driven 'odometer'. This discovery made the front cover of Science in 2000 and rewrote the classic, 1960s textbook notion of Nobel laureate Karl von Frisch that distance flown by a bee is estimated in terms of energy consumption.

Srinivasan's conclusion was based on his earlier observations of the visual system and navigation abilities of bees. His team at the ANU then proved his idea experimentally.

Bees travelling a short distance in a narrow tunnel were fooled into thinking they had gone a long way by varying the perceived motion of their surroundings.

"The closer an object is to the eye, the faster it seems to go past," he says. "It's like if you drive from Brisbane to Sydney in a car - the surroundings go past very fast because they are close and you get the impression you have moved a long distance. But if you go by plane, the ground is so far away that it hardly seems to move - you don't get the feeling you have moved very far. This is exactly what happens in bees."

Boy's own adventure

In his congress presentation, Srinivasan will also describe some of the applications of his research and why his discoveries have become the subject of considerable attention, and funding, from the US and Australian military. Many of his findings relating to vision and navigation are offering novel, computationally elegant solutions to problems in machine vision and robot navigation.

"Actually I didn't go looking for this - it sort of found me," he says. "We started playing around with automaton vision systems at the suggestion of defence funding bodies and because of my engineering and electronics background."

In fact, his team still does most of the prototype work for the robotics applications, with the testing room and equipment looking like a small (and not so small) boy's dream Mechano set.

His team's projects include a robot that can steer through cluttered environments, a camera that can give panoramic, insect-like vision to robots and surveillance cameras, an autonomous navigation system for helicopters, and a design for 'microflyers' for NASA.

The idea with this last one is that a mothercraft would land on some sort of inhospitable terrain like Mars or a minefield and release several small, relatively inexpensive, autonomous microflyers to survey the surrounding terrain.

Not forsaking his interest in basic biology, Srinivasan is most excited about his more recent and current work on bee emotion and cognition, and will summarise his findings in this area. "These insects may not be the simple, reflexive creatures that they were once assumed to be," he says.

Srinivasan is conducting experiments designed to test the bee's ability for tasks outside their expected behavioural repertoire. As with the low vision and processing work, Srinivasan has already made landmark discoveries in this area.

First, he showed that bees can learn rather general features of flowers and landmarks, such as colour, orientation and symmetry, and apply them to distinguish between objects not previously encountered (published in Nature in 1993, 1994).

In addition, bees exhibit "top-down" processing, meaning that they use prior knowledge to detect poorly visible or camouflaged objects. "Bees pre-trained to recognise uncamouflaged objects could be then trained to detect the same objects once camouflaged."

These findings have potential applications in the design of machine vision systems for surveillance.

Bee handedness

Srinivasan and his team more recently found showed that bees can learn to form complex associations and to acquire abstract concepts such as "sameness" and "difference" (Nature 2001).

Bees are also capable of associative recall, meaning that exposure to one stimulus, such as a scent, can trigger recall of associated stimulus, such as a colour, or even of a navigational route to a food location. This faculty is likely to aid a foraging honeybee in finding, memorising and reliably returning to a good source of food.

Srinivasan' team can train bees under laboratory conditions to choose a specific colour or particular visual pattern when they have experienced a certain scent. The application for this research is also far-reaching. "Using this principle, a colony of bees could potentially be trained to detect landmines using a pre-programmed scent," he says.

Finally, his laboratory reported last year that bees display lateralisation of brain function, revealing that invertebrates also show 'handedness'. According to Srinivasan, "all of these observations suggest that there is no hard dichotomy between invertebrates and vertebrates in the context of perception, learning and 'cognition'; and that brain size is not necessarily a reliable predictor of perceptual capacity."

Srinivasan is also interested in the 'softer' side of bees. "Bees may show frustration, anger, even joy," he says. "I want to determine if they really can experience emotion." This work has the potential to transform neurological principles of emotion and cognition.

Flight of the humble bee

Not a biologist by trade, Mandyam Srinivasan holds an undergraduate degree in electrical engineering and a Master's degree in electronics from India (the country's top graduate in both). Seeking to combine engineering and biology, he undertook a PhD at Yale University in the US working on visual processing in insect eyes, and then in 1985 accepted a research fellowship at the Australian National University (ANU). Srinivasan is now a world leader in visual processing in simple biological systems.

Not satisfied with multiple degrees, Srinivasan also gained a D.Sc. in neuroethology from the ANU in 1994. His research team (initially comprising just himself) has produced around 180 publications, including 21 in journals such as Nature, Science, and PNAS. In recognition of his scientific standing, Srinivasan was an inaugural Australian Federation Fellow in 2001 and is a Fellow of the Australian Academy of Science, the Royal Society of London and the Academy of Sciences for the Developing World.

After more than two decades with the ANU in Canberra, Srinivasan moved to the University of Queensland in January 2007 to take up a professorship in visual neuroscience at the Queensland Brain Institute. He is looking forward to the new opportunities and collaborations presented by his move to Brisbane and hopes to diversify his experiments into the fields of molecular biology and pharmacology.

Work at the University of Queensland on purpose-built facilities for analysing his 'flight of the humble bee' is almost complete.