The Vulcan mind meld is no longer the stuff of science fiction. Scientists have electronically linked the brains of pairs of rats, enabling one to apparently share information with the other. And unlike the Vulcan technique, there’s no need for close contact: Using an Internet connection, one rat in Brazil sent signals to the brain of a rat in North Carolina.
Similar experimental setups may help scientists probe how the brain’s circuitry incorporates information and perhaps even lead to some kind of organic computer of interconnected brains, says neuroscientist Miguel Nicolelis of Duke University, who led the study.
Other researchers are skeptical. The experiments, reported online February 28 in Scientific Reports, have a gee-whiz quality but they don’t expand science’s understanding of brain-body interactions and have little practical use, says neuroscientist Lee Miller of Northwestern University in Evanston, Ill.
Years of research into how the brain and body communicate have led to brain-machine interfaces such as artificial limbs that can be controlled by signals collected by tiny microchips implanted in the brain. Much research is also dedicated to the inverse — sending signals to the brain to achieve sensations in the body, as a cochlear implant can to overcome hearing loss. In the new study, the scientists put the two together. Using pairs of rats, the researchers extracted information from one rodent brain and sent information to the brain of another.
Groups of rats were trained to be either “encoders” or “decoders.” Encoders learned a simple task: When an LED light went on above one of two levers, the rat had to hit the lever associated with the light to receive a reward. Tiny chips implanted in each animal’s motor cortex, the part of the brain tasked with executing movements, recorded the nerve firings that corresponded with pressing the lever on the right. The average of these signals helped the scientists evaluate the signal provided by the encoders later in the experiments.
The decoder rats also did pre-experimental training: When they received a burst of stimulus from the brain chip and then pressed the lever on the right, they got a reward. When they received virtually no stimulus (just one brief pulse) and then pressed the left lever, they also got a reward.
Next the scientists connected each encoder rat to a decoder rat via wiring with a small computer in between. When an encoder rat pressed the lever on the right, the neural firing signal went to the computer, which sent the decoder rat a signal that was adjusted based on the average signal that had been calculated earlier. When an encoder pressed the lever on the left, the computer translated it into one or no pulse and sent that information to the decoder rat.
Encoder rats pressed the correct levers associated with the LED about 95 percent of the time. Each decoder rat, which didn’t have LED lights as a guide, received a signal from its brain chip, either the mathematically transformed signal for the right lever, or one or no pulse for the left. These decoder rats pressed the corresponding lever in 60 to 72 percent of the trials.
After the researchers began rewarding the encoder rats when the decoder hit the correct lever, the encoders’ behavior appeared to be influenced by the decoders’ activity. When a decoder rat hit the wrong lever, the encoder was quicker to hit a lever the next time around. And the encoder’s neural signal was stronger.
“When the decoder commits an error and the encoder doesn’t get a reward, it makes its behavior more accurate,” says Nicolelis.
The decoder rats did surpass the level of 50 percent correct responses that would be expected by chance. But the experiment would be much more interesting if the decoders hadn’t been trained to associate a burst of brain stimulation with one lever and a lack of stimulation with the other, says cognitive neuroscientist Uri Hasson of Princeton University. A rat might just be associating a brain zap with the lever on the right, and no zap with the lever on the left, as it was taught to during training, Hasson says.
