Copyright © 2007-2017 Russ Dewey
For most of the middle 20th Century, American psychologists paid little attention to classical conditioning, except for teaching students about Pavlov's dog. Then classical conditioning procedures starting showing up in neuroscience labs.
In the 1960s and 1970s scientists found that responses to classical conditioning could be measured in single neurons impaled with electrodes. By 2003, a single well-known neuron from the sea slug (Aplysia), called B51, could be conditioned in isolation (Mozzachiodi, Lechner, Baxter, and Byrne, 2003).
Neurons normally receive stimulation from other neurons. That is what activates incoming synapses, forming the biological basis of a US or CS.
In the dissected neurons from the Aplysia, the focus was upon conditioning in a single B51 neuron. Electrical stimulations substituted for the usual synaptic inputs.
After a few pairings of the electrical US and CS, the neuron exhibited enhanced output responses. In a whole animal, these would stimulate increased feeding responses. The clear implication is that classical conditioning is a very basic form of learning that does not even require a complete nervous system.
The Aplysia has about 20,000 neurons, but classical conditioning can be shown in a single neuron from its nervous system. That helps to explain why classical conditioning can be involved in all sorts of unconscious biological processes.
How was classical conditioning demonstrated in a single neuron?
Other researchers showed classical conditioning produces noticeable changes at the junction between neurons, the synapses. A synapse involved in successful conditioning became stronger, shown visibly by a thickening around the area of the synapse.
What changes were observed at synapses involved in conditioning?
Eric Kandel of Columbia University won a Nobel Prize in 2000 (shared with Arvid Carlsson and Paul Greengard) for his multi-decade research program on this and related topics. The main laboratory subject Kandel used was the Aplysia although later he extended his work to mice.
One form of unconscious learning that appears to be due to classical conditioning is drug tolerance Drug-taking behavior (such as using a needle or even opening a bottle of beer) functions as a signal or CS that predicts introduction of a drug into the body.
Eventually the act of drug taking triggers an anticipatory response: the secretion of drug antagonists that help eliminate the drug from the body. This lessens the body's response to the drug, which is what is called drug tolerance.
How does classical conditioning explain drug tolerance?
The ability of experienced drinkers to "hold their liquor" (consume a lot of alcohol without showing much effect) is a sign that the body is adapting to the drug. Classical conditioning has occurred.
Alcohol consumption now triggers a strong anti-drug action that reduces the effect of the drug. The same thing happens with cigarette smoking and coffee drinking. As a person becomes addicted, the drug has less and less effect.
Siegel and colleagues (1981) demonstrated that drug tolerance was due to classical conditioning. They gave rats morphine, a relative of heroin.
Sometimes the rats received a signal followed by a period of no morphine. Therefore the signal indicated a drug-free period. This can be called a C-, or "safety signal."
Gradually the rats built up a tolerance to the morphine. Its pain-killing effects disappeared.
But when morphine was given after the signal (typically a drug-free period) the rats had no tolerance for the morphine. The full pain-killing effects returned.
Intrigued by this finding, Siegel and co-workers (1982) gathered data on heroin overdose deaths in humans. They found that victims of heroin overdose were typically occasional or weekend users, not daily users.
Often the fatal overdose occurred in unfamiliar environments. Drawing an analogy to their rat studies, the researchers proposed that addicts who took the drug in an unfamiliar setting, or took it after a period of time not using the drug were in special danger of overdosing. Their bodies failed to perform the usual anticipatory response of secreting opiate antagonists, so they had less tolerance.
A potentially important application of Pavlovian conditioning involves the body's immune system. Like other body systems, it can be activated or suppressed through classical conditioning.
This has exciting implications. If learning can stimulate immune system activity, people should be able to arrange conditions to improve health or healing.
Perhaps humans have already been doing this for thousands of years. Classical conditioning might eventually shed a light on healing rituals practiced by pre-modern cultures.
The analogy to pain-reducing effects of placebo effects makes this plausible. Those effects were found to be due to endorphins, the body's endogenous opiates.
A stimulus associated with pain relief (and expected to provide pain relief) triggered the natural pain-relieving chemicals. Perhaps a stimulus associated with healing (and expected to provide healing) can do the same for natural immune responses.
How does the analogy to placebo pain relief make it "plausible" that ancient rituals might stimulate healing?
An early experiment reported by Schmeck (1985) involved a team of researchers at the University of Alabama medical school. They studied effects of classical conditioning on activity of natural killer cells (NK cells)
NK cells are lymphocytes: white blood cells that destroy germs and other invaders in the body. In the experiment, mice were exposed for three hours at a time to a powerful odor (camphor).
Exposure to this odor, by itself, had no effect on the mice. This had to be shown first, to rule out healing effects due to the camphor itself.
Next, mice were exposed to camphor odor then injected with poly I:C (polyinosinic-polycytidilic acid) which stimulates activity of natural killer cells. Mice in the control group did not receive poly I:C.
For the experimental group, the odor of camphor was paired with exposure to Poly I:C nine times. In the 10th session, the mice were exposed only to the odor of camphor. Every mouse in the experimental group showed large increases in natural killer cell activity.
Their bodies were "predicting" the injection of poly I:C and responding with immune system activity. In the control group, which was exposed only to the odor of camphor, no such response occurred.
How did researchers demonstrate a conditional immune response in mice?
If you are thinking this through, you might wonder, "Why did response to this poly I:C drug increase, when classical conditioning usually leads to a diminished responses to drugs, like with drug tolerance?"
However, in tolerance, the body prepares antagonists to an invading chemical, the drug. In this case, poly I:C is the invading chemical.
The body must interpret poly I:C as something to be fought off. When given a signal the poly I:C is coming, the immune response is triggered as an anticipatory response.
This shows what Rescorla meant when he said research on classical conditioning is "more than spit and twitches." Pavlovian conditioning can produce remarkable, subtle biological effects.
Analyzing the exact mechanism can be difficult, but it promises to be a fruitful investigation. How exactly does a mouse's expectation that poly I:C is about to be delivered to its bloodstream stimulate the production of NK cells?
What is the mechanism, on a neural level? If researchers knew that, perhaps they could help human patients boost production of NK cells when needed.
This type of finding led to a new discipline, neuropsychoimmunology, which promised insights into psychological effects on immune functioning. Results in the 20th Century were disappointing, however. Perhaps that field will blossom in the future as more knowledge is gained.
Schmeck, H. M. (1985, January 1). By training the brain, scientists find links to immune defenses. New York Times, pp.Y13-Y14
Siegel, S., Hinson, R. E., & Krank, M. (1981). Morphine-induced attenuation of morphine tolerance. Science, 212, 1533-1534.
Siegel, S., Hinson, R. E., Krank, M. D., & McCully, J. (1982). Heroin "overdose" death: Contribution of drug-associated environmental cues. Science, 23, 436-437.
Mozzachiodi, R., Lechner, H. A., Baxter, D. A., & Byrne, J. H. (2003) In vitro analog of classical conditioning of feeding behavior in aplysia. Learning and Memory, 10, 478-494.
Write to Dr. Dewey at email@example.com.
Copyright © 2007-2017 Russ Dewey