Gambling Behaviour in Pigeons: Toward an Animal Model of Gambling

Abstract

The gambling addiction is a destructive impulse control disorder that leaves families destitute and lives in ruins. The treatment and pathology of gambling has become a popular area of research, reflected in national media campaigns here in New Zealand. Studies suggest that there are prefrontal activation differences in problem gamblers compared to healthy volunteers, areas involved with decision making and reward. Animal research has begun to investigate an animal model of gambling, with some studies showing behavioural similarities between animals and humans when playing gambling tasks. The current thesis sought to further investigate an animal model of gambling, performing two experiments using a slot machine task. Experiment 1 was an investigation into the neural basis of gambling, examining the behaviour of single neurons in the avian NCL, an equivalent to the mammalian prefrontal cortex. Four pigeons (Columba livia) served as subjects and played a touch screen slot machine task, similar to a slot machine found in any casino. Pigeons were required to peck an upright arm to initiate each trial, and then peck four rolling tumblers in succession from left to right. If four identical stimuli appeared, then a wheat reward was won. During the task, activity from single NCL neurons was recorded. Four neuronal types were found to correlate with gambling related behaviour: Reward Proximity neurons, I-Won neurons, I-Lost neurons and Near Win neurons. In addition pigeons were split into two groups with one group trained for a short (one month) and long (four month) period, in an attempt to mimic naive and problem gamblers. It was hypothesised that due to prolonged training the two groups would differ in the amount and magnitude of firing of these neurons. Results showed no statistical difference between groups in either amount of gambling neurons or magnitude of firing. Behavioural data were also collected and it was hypothesised that all birds would show evidence of a post reinforcement pause, something our results confirmed. Experiment 2 investigated whether pigeons, similar to humans, show a preference to play slot machines with higher near win ratios. Four pigeons were again tested on a touch screen slot machine task, required to play on two different ‘machines’ differentiated by different coloured backgrounds. It was hypothesised that when pigeons were given a choice, they would choose the machine that had the higher percentage of near win trials. Our results did not show evidence of any preference, and it was concluded that the near win effect may not be replicable in animals. In conclusion, the current thesis demonstrates that neurons in the avian NCL code for gambling behaviours while pigeons play a slot machine task. Although an analogue of slot machine gambling may be found in animals, our findings suggest gambling characteristics such as the near win effect may not be replicable. The current findings provide a base for further research to investigate an animal model of gambling.