For the last 80 years, science has distinguished two forms of associative learning: classical and operant conditioning. Classical conditioning denotes the kind of learning made famous by the Russian physiologist I.P. Pavlov, who trained dogs to salivate in response to the tone of a bell by always feeding the animals after he rang the bell. About 20 years later, the American psychologist B.F. Skinner trained rats to press a lever for a reward – operant conditioning.
Until now, it was assumed that these and many other types of learning relied on a set of well-known learning-related genes. This assumption has now been questioned by the results from the scientists in Berlin. Brembs and his team at the Institute of Biology of Freie Universität Berlin studied the learning performance of genetically engineered fruit flies in a flight simulator.
In the first experiment, the cylindrical drum within which the fly was flying was illuminated alternatingly in blue and green. With, say, blue illumination the flies also received an aversive heat stimulus. This treatment resulted in an avoidance of the color blue.
In a second experiment, the colors and heat were coupled to the turning attempts of the fly such that the blue illumination and heat were always turned on when the fly was attempting to turn right. Thus, similar to Skinner's rats, the flies had the opportunity to actively learn to behave in a certain way – in this case to turn left to avoid being heated.
In a third experiment the scientists tested purely operant learning (behavioral learning). This was done by removing the color stimulus. The fly was still heated when it attempted to turn right, but now there were no colors presented any more.
Flies where the well-known group of learning genes had been manipulated failed miserably in the first two experiments. However, in the third set of experiments, they learned even better than normal flies.
So far it had been assumed that all forms of associative learning rely on the already known set of learning-related genes. According to Brembs, "Our studies show that the prominent learning genes are not playing any role in purely behavioral learning. Our results suggest that Pavlovian learning somehow suppresses behavioral learning."
Importantly, genetically engineered flies in which the enzyme protein kinase C was inhibited, showed an impairment in the purely behavioral third experiment, while they were normal in the other two learning tasks. Brembs posits that the operant, protein kinase C-dependent learning mechanism also underlies so-called "habit formation," which is implied in the acquisition of a drug-taking habit. If that were the case, the development of compulsive drug-taking could be slowed down or even prevented by medication interfering with the protein kinase C pathway.
Whether the same molecular learning mechanisms discovered by Brembs in flies also exist in mammals and possibly even in humans, is still unknown. "To find this out, one would need equivalent studies in mice or rats."
For further information, please contact:
Institute of Biology - Neurobiology
Freie Universität Berlin