Different Types Conditioning Methods: Classical And Operant Conditioning
This paper investigates the different types of conditioning methods that can be used on Siamese Fighting Fish, also known as the Betta Fish. Classical and Operant conditioning were used during this experiment to simulate different behavioral responses in the fish. A red Veiltail male was purchased in January of 2019 and is adapted to its new environment in about one week and after about 3 weeks of normal feeding and no training, the experimental training process began. The training exercises took place between January and April of 2019. Each training assignment was performed for 14 consecutive days and usually consisted of two to three trials per day, depending on how well the fish would respond that day. Each month consisted of different training tasks that included a visual stimulus with a marker tip, shaping through the use of a plastic hoop, an FR2 schedule to choose between two different colored markers, and finally a discriminative stimulus training session of acoustic cues, between two different genres of music.
The number of successful, unsuccessful, and non-responsive trials were recorded for each train session and this is what each training session was centered around.
The Conditioning of Betta Fish Through Classical and Operant Conditioning
Siamese Fighting Fish or Betta Fish are admired and accepted fish in the Aquatic World. Bettas are known for their vibrant colors and exotic tails. Although these fish are relatively simple to take care of, they are unable to cohabitate with any other Betta fish due to their aggressive and territorial behavior.
Several studies have been conducted on Betta fish, in terms of operant and classical conditioning. Classical Conditioning is defined as, “a process whereby a stimulus comes to elicit a response because it has been paired with or associated with another stimulus.” (Powell, Honey & Symbaluk. 2017, p. 546) An experiment was conducted on the aggressiveness of male Siamese fighting fish. Betta’s are known to be aggressive fish so, in order to prove this, betta fish was presented with a mirror (US) and the (UR) would be to attack due to Betta’s feeling like they are threatened. Through several repeated trials, the (UR) became the (CR) with the help of classical conditioning and the pairing of associating a mirror with an intruder. (Thompson & Sturn, 1965) Involved with the same study, researchers also explored gill reaction patterns to an electrical shock. The electrical shock (CS) was paired and presented concurrently with a mirror (US) in order to provoke an aggressive response in the Betta. The shock caused the gills to flare out. As stated previously, Betta’s will react if they see their reflection or if there is an intruder in the tank. After repeated trials, the CS lead to a CR of the gills of the fish to flare, due to it associating the shock with the mirror (Thompson & Sturn, 1965).
Operant Conditioning differs from Classical Conditioning. Operant Conditioning is defined as, “a type of learning in which the future probability of a behavior is affected by its consequences”. Similar to Thorndike’s, Law of Effect.” (Powell, Honey & Symbaluk. 2017, p. 215) Thorndike’s Law of Effect believes that activities that are rewarded will most likely be repeated and the opposite also holds true for activities that lead to punishment, they will not be repeated. Stimuli that provoke normal aggressive behavior in a betta can also serve as a reinforcer for operant conditioning. (Thompson & Sturm, 1965) In a similar study, three different colored Betta fish were presented with a visual stimulus and each fish had to swim through three gates in order to be presented with a mirage of a different colored Betta. Six varying colors of Betta’s were presented to the fish. The mirage of the fish traveled across the tank from left to right and this is what provoked the aggressive display in each Betta. However, the mirage was then removed, and the Betta fish had to backtrack through the three gates in order to seek the mirage again. This study demonstrates how Betta’s are able to discriminate between colors. Betta A responded the most to green fish, whereas Betta B & C had the greatest recurrence rate of response to the redfish. Thompson and Sturm found that “the data here suggests that the subject’s coloration with respect to the model’s coloration may be an important variable determining operant reinforcing strength.” (Thompson & Sturm, 1965)
Both studies focused on the theme of visual cues and it is recognized that through the use of visual cues, Betta fish are able to learn through both classical and operant conditioning. For this experiment, four hypotheses were theorized for each section of the experiment. The assignment I was that a visual cue of a red expo marker top will be paired with a food pellet to encourage the behavior of the Betta to swim to the top for food. Assignment II involved operant conditioning in the sense that it would elicit the behavior of the fish swimming through the hoop for a food pellet. Assignment III trusted that an FR-2 schedule would promote the behavior of the fish to successfully react to the same marker twice in order to obtain a food pellet. Lastly, assignment IV expected that through the use of discriminating training, the betta would be able to discriminate between two different genres of song and would receive a food pellet.
One red Veiltale Male Betta fish (age unknown) was used for this experiment. It was purchased from PetSmart in North Bergen, New Jersey. The fish was purchased in January of 2019.
All materials were purchased from PetSmart in North Bergen, NJ. The Betta Fish habituated in an Imagitarium Betta Desktop Kit Tank (1.4 gallons). The Tank included 2 bags of colored glass pebbles, a pineapple, and a water heater. The tank was at 72-73° F from February to the end of the experiment and training blocks. During January, the temperature of the tank dropped to 67-69°. Hikari Bio-Gold Betta fish food (.088 oz) and Top Fin Betta Bits Color Enhancing Pellets Fish Food (1 oz) was used as well as Top Fin Betta Water Conditioner (4 fl oz). Materials that were used during the training blocks were two sharpie markers (red and yellow), a red expo marker, a plastic hoop, MacBook Pro for music, and an iPhone XS Max for video recording purposes.
After three weeks of acclimation, the training sessions began. Training Session I continued for 14 days and each day consisted of two to three trials, depending on how well the fish responded to the foreign stimuli. Training Session I consisted of a red expo marker being inserted into the tank. Each time the Betta fish would recognize and react to the marker, it would receive a food pellet. The US was the food pellet that caused the UR of the fish-eating it. To the fish, the marker was an NS because it did not elicit any response. Through the repeated trials of classical conditioning, the NS became the CS which in turn caused the fish to swim to the top and eat. The CS lead to a CR of the fish receiving a pellet every time it recognized the marker. The fish paired the marker with food and would swim to the top of the tank each time it recognized the marker.
Training Session II was also 14 days long and consisted of two to three trials per day. This activity comprised another visual cue. The Betta fish had to swim through a plastic hoop and once it completed this, it would receive a food pellet. A food pellet was inserted on the other side of the hoop. The SD was the plastic hoop and the R was the fish swimming through it. The SR was the pellet being received once the fish successfully swam through the hoop.
Training Session III was based on Fixed-Ratio (FR2) schedule. This session was 14 days long with two to three trials per day. For this activity, two sharpie caps were inserted into the tank. The betta fish had to “flare it gills” or simply just react twice to the same marker and once it did this successfully, it would receive a food pellet. The aggressive behavior of the betta was used to the experiment’s advantage.
Training Session IV was centered around discriminative stimulus training. This training lasted for 14 days and consisted of two to three trials per day. Through the use of classical conditioning, the fish had to discriminate between two different types of music. The song I was an upbeat, rock song by AC/DC, and Song II was a mellow, slow song by Khalid. The song I was the CS for a food pellet whereas song II was the US. The fish would respond solely to Song I.
Throughout the four training activities, both classical and operant conditioning was used in order to accomplish successful training of the Betta fish.
For the training activity, I, it was hypothesized that through the use of classical conditioning, the tip of the marker would cause the fish to swim to the top and then receive a food pellet. A pellet was given to the fish when he successfully noticed the marker and swam to the top of the tank. Training for this activity was 14 days long and there were 15 successful trials, 5 unsuccessful trials, and 8 non-responsive trials. Therefore, this supported the hypothesis that a visual cue, such as the expo marker tip is able to condition a fish through classical conditioning.
For training activity II, the idea of shaping was used here in order to reinforce a required behavior from the fish. It was hypothesized that through shaping, the fish would swim through a hoop. To begin this activity, I would insert the expo marker tip on the other side of the plastic hoop so that the fish would still recognize the expo marker and pair it with receiving food. Through repeated trials, the fish successfully swam through the hoop 12 times, there was 9 unsuccessful trials and 7 non-responsive trials. This too supported the second hypothesis that through shaping, the fish was able to successfully swim through the plastic hoop.
For training activity III, it was hypothesized that an FR2 schedule would allow the fish to achieve the desired behavior of receiving a pellet. The Fixed-Ratio aspect enforces the behavior of the fish by successfully completing an action twice before any reward was given. For this activity, the Betta fish had to successfully choose between two different colored markers and once completed, he had to react twice to one single marker by flaring it gills or swimming close to the desired marker. The reinforcer for this activity was the food. For this training assignment, there was 4 successful trials, 12 unsuccessful trials and 6 non-responsive trials. The reason as to
Lastly, training activity IV used discriminative training. The hypothesis for this activity was that the fish had to discriminate between two different types of music and once this was completed, it would receive a food pellet. The betta fish had to discriminate between a fast rock song or a slow-paced, more mellow song. The betta fish would receive a food pellet when it would respond and swim to the top when the rock song was played. When the slow, mellow song was on, the fish would not receive a pellet of food. The discriminative stimulus for this activity was the AC/DC rock song. For this activity, there were 4 successful trials, 5 unsuccessful trials and 7 non-responsive trials.
The chart above demonstrates the variability of behavioral responses from the Betta fish. Each assignment is divided into successful, unsuccessful and non-responsive trials. The assignment I and II had the best outcomes in terms of successful attempts. However, assignment III and IV has the most variability in terms of success rate. During this time is when the Betta fish developed swim bladder disease which caused the fish to be extremely lethargic. This may have possibly caused the decline in successful attempts. The assignment I and II had exactly 28 trials over a two-week period. However, assignments III and IV had about 20-25 trials due to the illness of the Betta fish.
For the first two assignments, the hypotheses were supported and proven correct. The visual cue for an assignment I was able to successfully train the fish to pair the marker tip (NS) with a food reward. The (NS) became the (CS) over time. The CR was the fish swimming to the top for a food pellet. With Thompson and Strums first experiment, they successfully were able to pair a mirror (US) with the Betta’s fish’s aggressive behavior (UR). Although these two experiments were different, they yielded similar results. For Thompson and Strums experiment, the mirror served as (CS) which brought out aggression in the fish (CR). For this study, the food served as (US), the expo marker tip (CS) caused the (CR) of the Betta fish to swim to the top for a food pellet.
In terms of assignment II, shaping was used here. As stated previously, this relates to the Law of Effect. Each time the fish would swim through the hoop, it was given a food pellet. this means that the likelihood of receiving a food pellet is greater as the fish continues to swim through the hoop.
For assignments III and IV, this is where the results skewed and greatly decreased. During this time, the Betta fish used for this experiment developed Swim Bladder disease. The swim bladder is described as “an area to compensate for downdrift caused by the high densities of the skeleton and muscles. The bladder itself is used by the fish to either rise or sink in the water” (Fänge, 1983) However, due to several trials of repeated feeding, the swim bladder became enlarged due to constant eating and no defecation. The fish became constipated and would lay at the bottom of the tank. Rarely, it would rapidly swim to the top for air and then drop back down to the tank for hours at a time. It would constantly hide and would not nearly be as responsive as it usually was. Besides remaining only at the bottom of the tank, the fish also began swimming on an angle due to increased inflammation in its belly. However, I was able to do some research and it said to not feed the fish for three days. This should allow the fish to defecate while also not receiving any food. Due to this, the results for assignments III and IV were so low. It was difficult to get the attention of the fish and then to have it interact with the several stimuli in the tank. However, for assignment IV, the fish was not completely healed but it was somewhat more active, and it was able to discriminate between the two genres of music correctly four times. Similar to the experiment by Thompson and Strum, the three-betta fish were able to discriminate between six different colored mirage betta’s, and depending on the betta fish’s color, it had the highest aggression rate that was consistent towards a certain mirage color.
Some limitations for this experiment that I noticed were that the original placement of my fish tank was detrimental to the fish’s health. The fish was purchased in January and the tank was located near a window. The draft from the window caused the temperature to drop and changed the behavior of the fish for some time; this is why a tank heater was inserted into the tank during late February and the tank was also moved away from the window. I am unaware if this counts as a limitation but the food for the Betta fish was changed halfway through the experiment. This was not done purposely; the food ran out and the previous food was not in stock at PetSmart. Another limitation that I noticed was that the fish was a responsive fish to begin with and every time I or other people would walk past the tank, it would immediately swim to the top of the tank looking for food. I do not believe these would have skewed any results of the experiment, if anything, it may have helped.
I thoroughly enjoyed this project. My father and I have always been fishing people but I was never given the opportunity to train a fish before. I have had betta fish since my freshman year of college and realized very quickly that they are smart and interactive fish. I know it’s easy to train a dog or even a cat, but I never thought a fish could be trained. I would not change anything about this project, I think every aspect of the project was straightforward forward and the instructions were descriptive and not too difficult to perform. Through the class and this project specifically, I learned a lot about both humans and animals. Specifically, I learned how fish can be trained and their behaviors manipulated through the use of classical and operant conditioning.
- Powell, R. A., Honey, P. L., & Symbaluk, D. G. (2017). Introduction to learning and behavior. Boston: Cengage Learning.
- Thompson, T., & Sturm, T. (1965). Classical conditioning of aggressive display in Siamese fighting fish1. Journal of the Experimental Analysis of Behavior,8(6), 397-403. doi:10.1901/jeab.1965.8-397
- Fänge, R. (1983). Gas exchange in the fish swim bladder. Reviews of Physiology, Biochemistry and Pharmacology Reviews of Physiology, Biochemistry, and Pharmacology, Volume 94,111-158. doi:10.1007/bfb0035347
- Thompson, T., & Sturm, T. (1965, September). Visual-Reinforcer Color, And Operant Behavior In Siamese Fighting Fish. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1338109/