Nature's amazing synchrony explained

November 15, 2011 - via The Age

A team of biologists from Australia and Sweden has studies schools of fish to better understand their movement.
After being mesmerised by a school of fish swimming gracefully in formation, then dispersing and effortlessly reforming, you may have been reminded of ballet dancers moving in exquisite synchrony. But fish motions, though clearly co-ordinated, are not choreographed. So how do they do it?

Biologists are now closer to an answer. A team from Australia and Sweden studied shoals of mosquito fish to determine if they use the rules set by mathematical models. Introduced into Australia in 1925, the fish — known scientifically as Gambusia holbrooki — are now considered noxious pests.  After filming groups of two, four or eight fish for five minutes, semi-automated tracking software was used to analyse the motion patterns and trajectories of each fish.

To the biologists’ surprise, each was found to use a relatively simple set of rules to respond to its neighbours. ‘‘They use some of the rules of models, but don’t use others,’’ lead researcher James Herbert-Read, of Sydney University’s school of biological sciences, says. ‘‘It turned out to be rather like driving a car,’’ he explains. ‘‘The fish mainly use visual cues to see where other group members are, and then adjust their behaviour depending on their neighbours’ positions.’’

The rules, he says, include accelerating towards neighbours that are far away and decelerating when neighbours are right in front. In other words, fish respond to the position of neighbours through rules relating to short-range repulsion and longer-range attraction. ‘‘We also found that a fish only responds to a single nearest neighbour at any one time.’’

The research, published in the latest issue of the  American journal Proceedings of the National Academy of Sciences, along with another recent work, is the first demonstration of rules that individual fish use to interact in shoals.

The rules drive the processes underlying swarm intelligence, a type of intelligence based on the collective behaviour of self-organised systems. These are collections of small, interdependent units that spontaneously form organised structures or patterns without external intervention by a central authority.

‘‘Groups are better at gathering information than singletons,’’ Mr Herbert-Read says. But this information needs to be distributed among group members if it is to be useful to all individuals. ‘‘The rules we have identified drive this transfer of information.’’

How they do it

Individual fish use the position of neighbours to adjust their behaviour. So if one fish was scared by a predator, it would be expected to move away from the aggressor. Then the others, using these simple rules, would follow suit. ‘‘In this way, information is passed between group members using the rules we have identified,’’ Mr Herbert-Read says. The findings also explain how other animal groups move in a coherent and co-ordinated fashion, including the way flocks of birds fly in formation and herds of cattle or horses hurry along in harmony.

‘‘So, in future, when other researchers try to identify rules of interactions in other species, they may use our study as [a] starting point,’’ Mr Herbert-Read says. ‘‘It may be that different species use similar rules and this would go to show that these rules are incredibly efficient in driving co-ordinated group movement in different animals.’’

Decentralisation
Certain self-organising principles lie at the root of these behaviours, the scientists believe.  No one individual in the group always controls what the group does, Mr Herbert-Read says. ‘‘This is what we call ‘decentralised control’.’’ Other examples include web search engines and the sharemarket.
A question of anarchy
Might decentralised control in the animal kingdom be an example of leaderless behaviour? ‘‘It may be that one fish initiates a movement at a particular instance, but then another takes over, and so on,’’ Mr Herbert-Read says. ‘‘There are no consistent leaders in these groups — even though there are instances of leadership — and leadership, when it emerges, is distributed among group members over time.’’

The scientists are now investigating which individuals are initiating changes in direction. ‘‘We need to have individuals that have information and therefore lead, and others that don’t have information and therefore follow,’’ Mr Herbert-Read says. This can be done by training fish to associate particular stimuli with rewards, and not training others.

In some ways, the collective motion of animals is reminiscent of what happens in a branch of physics called fluid dynamics. It also seems to be related to an aspect of chaos theory, known as emergent behaviour — in a nutshell, self-organisation produced from local rules that determine how individuals interact.

‘‘Both systems, therefore, are the emergent outcome of the way individuals interact with one another,’’ he says. ‘‘Fluids move according to rules governed by the laws of physics, while fish also respond to behavioural rules.’’

Group dynamics
‘‘Biologists have long appreciated the social significance of group behaviour in the animal kingdom, among which they recognise the co-ordination exhibited by creatures such as ants and bees — so-called social insects — as something special and remarkable,’’ writes Philip Ball in his book Flow.  ‘‘But only rather recently have these collective motions been seen as something akin to flow.’’

A member of Mr Herbert-Read’s team, Ashley Ward, also from Sydney University, believes  the rules of interaction of most grouping species — from insects to people and all creatures in between — are likely to be based on similar principles.

‘‘The fish system is the first to be mathematically determined,’’ Associate Professor Ward says. ‘‘This is important because a massive number of species form into groups, including most of the animals that we commercially exploit — plus, of course, ourselves, when humans form into crowds.’’

The challenge now, he believes, is to empirically determine the rules of interaction in other grouping species and to find the similarities and differences.

‘‘We expect there to be similarities in the rules used by different species,’’  Professor Ward predicts. ‘‘It’s likely, therefore, that the rules used by fish in this study are not a million miles from the rules that you and I use as we interact with people in a human crowd.’’

Future directions
Technologists and engineers are likely to learn from the fact that individuals in animal groups get really close to one another and yet still avoid congestion and collisions.

‘‘Engineers have the potential to program some of the rules we have identified into robotic systems to develop transport systems that avoid everyday human traffic problems,’’ Mr Herbert-Read says. ‘‘The key here is that individual components of the system would need to interact with one another for this to work.’’

Roboticists would also be potential beneficiaries. Tiny robots, for example, can only achieve so much: they are poorer in sensing and communicating motion and lack the processing power and instrumentation of their bigger cousins. But if micro-robots can work in a co-operative manner, their performance may be comparable or even superior to that of a single bulky bot.

To work co-operatively, future baby bots might rely on swarm intelligence. Robot researchers, who have already studied ant colonies, bees and bird swarms, have developed algorithms for co-ordination and self-assembly. The new research is likely to spur another generation of robotically synchronised movers and shakers.
LINKS
    •    Find out more about swarm intelligence and related matters in the book Flow (Oxford University Press).
    •    Learn about emergent behaviour in biological systems at:www.complexity.org.au/ci/vol01/green01/green01.pdf
    •    Discover the magic of self-organisation at:epress.anu.edu.au/info_systems/mobile_devices/ch09s05.html

External link: http://www.theage.com.au/world/science/natures-amazing-synchrony-explained-20111115-1ngiq.html

Tags:
Author:Peter Spinks

Search Community News

Browse Archive

Top Stories of the Months

Categories