IMAGINE getting inside the mind of a shark: swimming silently through the ocean, sensing faint electrical fields, homing in on the trace of a scent, and navigating through the featureless depths for hour after hour.
We may soon be able to do just that via electrical probes in the shark’s brain. Engineers funded by the US military have created a neural implant designed to enable a shark’s brain signals to be manipulated remotely, controlling the animal’s movements, and perhaps even decoding what it is feeling.
That team is among a number of groups around the world that have gained approval to develop implants that can monitor and influence the behaviour of animals, from sharks and tuna to rats and monkeys. These researchers hope such implants will improve our understanding of how the animals interact with their environment, as well as boosting research into tackling human paralysis.
More controversially, the Pentagon hopes to exploit sharks’ natural ability to glide quietly through the water, sense delicate electrical gradients and follow chemical trails. By remotely guiding the sharks’ movements, they hope to transform the animals into stealth spies, perhaps capable of following vessels without being spotted. The project, funded by the Defense Advanced Research Projects Agency (DARPA), based in Arlington, Virginia, was presented at the Ocean Sciences Meeting in Honolulu, Hawaii, last week.
Neural implants consist of a series of electrodes that are embedded into the animal’s brain, which can then be used to stimulate various functional areas. Biologist Jelle Atema of Boston University and his students are using them to “steer” spiny dogfish in a tank via a phantom odour. As the dogfish swims about, the researchers beam a radio signal from a laptop to an antenna attached to the fish at one end and sticking up out of the water at the other. The electrodes then stimulate either the right or left of the olfactory centre, the area of the brain dedicated to smell. The fish flicks round to the corresponding side in response to the signal, as if it has caught a whiff of an interesting smell: the stronger the signal, the more sharply it turns.
The team is not the first to attempt to control animals in this way. John Chapin of the State University of New York Health Science Center in Brooklyn has used a similar tactic to guide rats through rubble piles (New Scientist, 25 September 2004, p 21). Chapin’s implant stimulates a part of the brain that is wired to their whiskers, so the rats instinctively turn toward the tickled side to see what has brushed by. Chapin rewards that response by stimulating a pleasure centre in the rats’ brains. Using this reward process, he has trained the rodents to pause for 10 seconds when they smell a target chemical such as RDX, a component of plastic explosives.
The New York Police Department is considering recruiting Chapin’s rats to its disaster response team, where they could be used to detect bombs or even trapped people, and Chapin met them to discuss the possibility last month.
However, Chapin’s “mind patch” only works in one direction: he can stimulate movement or reward an action, but he cannot directly measure what the rat smells, which is why he has to train them to reveal what they are sensing. DARPA’s shark researchers, in contrast, want to use their implant to detect and decipher the different patterns of neural activity that indicate the animal has detected an ocean current, a scent or an electrical field. The implant sports a small pincushion of wires that sink into the brain to record activity from many neurons at once. The team plans to program a microprocessor to recognise which patterns of brain activity correlate with which scents.
Atema plans to use the implants to study how sharks track chemical trails. We know that sharks have an extremely acute sense of smell, but exactly how the animals deploy that sense in the wild has so far been a matter of conjecture. Neural implants could change all that. “You get much better information from a swimming shark than from an anaesthetised animal that is strapped down,” says Atema. “It could open up a whole new window into how these animals interact with their world.”
At the Hawaii Institute of Marine Biology, Tim Tricas is using the implant to investigate what information scalloped hammerhead sharks glean from their electric field sensors. Gel-filled pores, scattered across a shark’s head connect to nerve endings that make them sensitive to voltage gradients. Sharks can use these electroreceptors to spot the weak bioelectric fields around hidden prey, such as a flounder buried in sand.
For decades, marine biologists have suspected that sharks might also use these electroreceptors for navigation. Tiger and blue sharks can swim mile after mile in a straight line with no view of the ocean floor and only scattered, changing light coming from above. Some researchers suspect they maintain their heading by using the Earth’s magnetic field.
When a conductor – in this case the shark – passes through a magnetic field, the interaction sets up a voltage across the conductor. The strength and orientation of that voltage depends on the conductor’s angle to the magnetic field. If a shark could detect those changes, it could use its electrical receptors like a compass. The only way to test this, Tricas says, is to monitor electroreception in a freely swimming shark.
Other animal behaviour researchers are setting their subjects loose too. Jaideep Mavoori at the University of Washington in Seattle has developed a neural implant for monkeys that can monitor brain activity while the primates play. “We believe we are the first to record neural activity from a monkey doing a somersault,” Mavoori says.
Mavoori’s implant can also stimulate one part of the brain in response to activity in another, and has a microchip that can interpret the neural signals and send a message to another part of the brain or a muscle accordingly. He and his colleagues believe such an implant might ultimately help humans compensate for lost nerve function caused by injury or disease.
They have found that when a monkey is free to move around, sets of neurons controlling opposing muscle groups – those that extend and flex a joint – are both active throughout many movements. However, when a monkey is restrained in a chair and taught to extend its hand for a food reward, say, only the neurons that control the extensor muscles tend to be active.
â€œRemote controlled sharks glide silently though the water without being spottedâ€?Understanding this difference may be vital in creating a muscle-stimulating prosthesis to restore movement to a limb paralysed by nerve damage. For some loose movements, such as gently extending your arm in and out, sending signals to opposing muscles in turn works quite well. However, for movements that require some rigidity in the joint, such as inserting a book into a bookcase, you need to engage opposing muscles simultaneously. A successful neural prosthesis will need to mimic both patterns.
Meanwhile DARPA too plans to take its shark implants out of the laboratory. Project engineer Walter Gomes of the Naval Undersea Warfare Center in Newport, Rhode Island, says the team’s next step will be to implant the device into blue sharks and release them into the ocean off the coast of Florida.
However, the radio signals used to direct the dogfish in the tank will not penetrate water, so the engineers plan to communicate with the sharks using sonar. According to Gomes, the navy already has acoustic signalling towers in the area that are suitable for relaying messages from a ship to a shark up to 300 kilometres away. The team has designed a sonar receiver shaped like a remora fish to minimise drag when attached to the animal.
The scientists will be particularly interested in the sharks’ health during the tests. As wild predators, it is very easy to exhaust them, and this will place strict limits on how long the researchers can control their movements in any one session without harming them. Despite this limitation, though, remote controlled sharks do have advantages that robotic underwater surveillance vehicles just cannot match: they are silent, and they power themselves.
From issue 2541 of New Scientist magazine, 01 March 2006, page 30
Fish with chips stay close to the farm
Fisheries scientists are investigating the use of neural implants to control the behaviour of farmed fish. They hope the tags will eliminate the need to pen and feed fish, a practice that pollutes the surrounding waters and promotes disease. Instead, the plan is to let the fish loose to forage for themselves and then retrieve them when they are large enough to harvest.
One way to contain the fish would be an acoustic fence, a barrier of sound signals that would trigger the implants to deliver a warning signal to the fish’s brain, possibly by mimicking a bad smell. Barry Costa-Pierce, a marine researcher at the University of Rhode Island in Narragansett, says his team has already developed implants that can make the fish surface on command. The project is focusing on bluefin and bigeye tuna, cobia and salmon.
Costa-Pierce is hoping to reduce the cost of each implant to a matter of pennies, although he admits the barriers to implementing the scheme are primarily legal, not economic. Setting tuna loose would raise the question of who owns a fish that swims in the commons of the ocean. Until governments can establish fishing regulations that take account of such implants, commercial fisheries are unlikely to take up the idea.