When we were first married, my husband and I frequented a local BBQ restaurant. He ordered the fried chicken – all drumsticks – and I loved the brisket sandwich. One weekend, we both developed tummy troubles the morning after we ate there, and automatically thought “It must have been the BBQ food!” My husband and I had very different responses. I refused to even enter the restaurant for two years as all reminders (the sign, the name, the lighting, the smell) made me nauseous. My husband, however, continued to eat his favorite fried chicken, despite getting sick two more times within two weeks of the initial incident. What had happened?
My avoidance response to any reminders of the BBQ incident carried two hallmarks of taste and context aversion learning, which differ from traditional classical conditioning. First, despite the long delay between eating the food and experiencing the gastrointestinal malaise, I associated the two events. Second, it only took one bad morning of digestive discomfort to turn me away from my favorite BBQ. Traditional classical conditioning relies on repeatedly pairing a neutral stimulus, like a bell, with an unconditioned stimulus, such as food, in close temporal proximity. In contrast, taste aversion can be acquired in a single trial.
Like Pavlov, who did not set out to research classical conditioning, it was mixture of astute observation and curiosity that brought John Garcia to the study of taste aversion. While preparing rats for the Radiological Defense Laboratory in San Francisco, he noticed a change in their drinking habits. Rats drank less water from plastic bottles, present in the cages where they received radiation, but showed no change in the amount they consumed from the glass bottles in their home cages. Garcia, who disliked the taste of coffee from plastic cups, wondered if the rats had associated the plastic-water taste with the radiation-induced nausea. As an informal test of his idea, Garcia switched the bottle locations and found that the rats continued to avoid drinking from the plastic bottles, even when they were placed in their home cages. This simple exploration was followed by a battery of controlled experimental trials on taste aversion learning in rats.
Ultimately, Garcia and his colleagues found that the pairing of taste and nausea produced the exceptional results. Conditioned taste aversion could develop after only a single exposure; it was robust to delays between the two stimuli, and lasted for long periods of time (just like my BBQ reaction). Other pairings, such as taste and an electric shock, bright lights and nausea, or environmental cues and radiation-induced illness, required repeated exposure and did not endure over time in the same way as conditioned taste aversion. Garcia and his colleague Koelling theorized that animals are “prepared” by nature to pair certain stimuli, such as smell and taste, where characteristics of the conditioned stimulus are strongly related to the unconditioned stimulus, or the reinforcer. The primacy of taste aversion has not gone undisputed, but proves difficult to test.
In a recent article in the Psychonomic Society’s journal Learning and Behavior, Sezen Kislal and David Blizard of Pennsylvania State University conducted a series of experiments to explore the nature of conditioned context aversion and to assess Garcia’s speculation that rats had learned to avoid the taste of water from plastic bottles. Conditioned context aversion – such as my reluctance to revisit the local BBQ restaurant – occurs when environmental stimuli are associated with the onset of illness. Kislal and Blizard, using specifically bred mice (a species that has not been tested for conditioned taste aversions often), performed three tests in which visual contextual cues were altered to explore the parameters of conditioned context aversion. All tests were performed within the home cage of the mice to facilitate the detection of a novel stimulus. Mice in the experimental group received an injection of Lithium chloride (LiCl), to induce nausea, and mice in the control group received an injection of sodium chloride (NaCl). The table below summarizes the experimental conditions.
The first test manipulated the type of bottle present in the environment and the feeling of queasiness. Reversing the procedure used by Garcia, Kislal and Blizard paired the LiCl with a glass bottle and the control with the plastic bottle. To directly test Garcia’s informal hypothesis that it was the plastic taste of water that became associated with the feeling of sickness from radiation, Kislal and Blizard included a plastic taste control group. They filled the control group’s glass bottle with water decanted from a plastic bottle. The rats were given three conditioning trials followed by a retention test trial 1 week later.
The mice showed an aversion to the glass bottles by the second and third trials (the black bars and the white bars in the figure below). This pattern was maintained after 1 week with no additional conditioning trials. The figure below illustrates these results. Strikingly, the plastic-tasting glass bottle group (white bars) was not significantly different from the other glass bottle group (black bars).
The second test manipulated the contextual cue so that either a large difference between the conditioned stimulus (graduated tube with dark tape) and the neutral stimulus (glass bottle) existed or a small difference existed (i.e., different colored tape) between the conditioned stimulus and the neutral stimulus. Once again, the mice drank significantly less water when the dark-tape tube was paired with LiCl than when the glass bottle was paired with NaCl, after only one trial.
When the contextual difference was less apparent, by contrast, two conditioning trials (CCA3—white bar in the figure below) were needed before a strong aversion was observed. These conditioned aversions were maintained for up to 13 days after the last conditioning trial (Second Retention—CCA3+13 days – black bars, white bars), suggesting that conditioned context aversions were strong, long-lasting associations.
The final experiment replicated the second experiment but inserted a 30 min delay between the presentation of the conditioned stimulus (a graduated tube with dark tape, light tape, or a glass bottle) and the unconditioned stimulus (a feeling of nausea from the injection of LiCl). Retention was tested at 7-day intervals for 3 weeks. As in Experiment 2, the mice needed two conditioning trials before displaying an aversion (CCA 3 in figure below). The effect was stronger when the LiCl was presented immediately after the CS (black bar) rather than after a 30 min delay (white bar). The effects for immediate and delayed presentations lasted a week, with no evidence for a continued aversion at either a two-week or a three-week follow-up. The mice also drank from the control bottle less in this last experiment, suggesting that the mice may have become sensitized to other contextual clues, like the presence of the experimenter, or procedural cues, like the water deprivation procedure used in the study.
In summary, Kislal and Blizard found that conditioned context aversion – like conditioned taste aversion – could occur after a single pairing, be formed across time delays, and have a long retention. Minor context changes in an otherwise familiar environment quickly became associated with nausea. Additionally, a plastic taste paired with a context change was not notably different from the context change alone. Kislal and Blizard argue that Garcia most likely observed conditioned context aversion learning rather than conditioned taste aversion in his initial discovery. Generally, Kislal and Blizard’s results support a broad perspective to studying aversion learning, less focused on the distinctiveness of taste.
Why should we care about the ability of mice to be conditioned by the context in which they feel nauseous? Conditioned aversions serve an important protective function, even in humans, because they discourage us from eating dangerous foods or standing near noxious fumes. Psychologists have even used conditioned taste aversion to keep coyotes away from livestock.
Unfortunately though, conditioned aversions are not always beneficial. Patients experiencing chemotherapy or radiation treatment experience nausea and suppressed appetites upon subsequent arrivals at the clinic following their first treatment, responding to contextual stimuli as they prepare for their next round of treatment. These patients also experience a reduction in immune functioning that is both a byproduct of the medicine used for treatment and a conditioned response to the environment. Understanding the contexts in which conditioned aversions may develop provides opportunities to facilitate survival while also improving the lives of others, whether it is a human, a sheep, a coyote, a rat, or a mouse.
As for my husband’s immunity to either context or taste aversion conditioning, it remains a mystery. Bon appetite!
Article focused on in this post:
Kislal, S., & Blizard, D. A. (2016). Conditioned context aversion learning in the laboratory mouse. Learning & behavior, 1-11. DOI: 10.3758/s13420-016-0217-2.
Note. This blog was written with contributions from Sarah Dietrich.