Dopamine and learning in the pigeon (columba livia)

Abstract

The direct dopaminergic agonist apomorphine (Apo) induces bouts of repetitive stereotyped pecking in the domestic pigeon (Columba livia). When an effective Apo dose is repeatedly injected, the total number of pecks elicited by every injection gradually sensitizes up to a dose dependent asymptote. Such response sensitization is a consistent model, which may constitute a valuable tool to study the involvement of the central dopaminergic system in learning. However, it was first necessary to investigate to what extent this well-known behavioral paradigm is based on learning or whether it includes non-learning components. A brief theoretical introduction is offered and the Apo-induced pecking sensitization in pigeons characterized. The first and most intuitive explanation for this phenomenon −physical accumulation of Apo in the nervous system− is ruled since longer inter-injection intervals do not reduce sensitization. Two further hypotheses remain: the "pharmacological hypothesis” states that sensitization results from Apo-induced pharmacological changes directly modifying relevant synapses; the "conditioning hypothesis" states that circulating Apo initiates learning processes that in turn modify synapses (in terms of learning, repeated Apo injections in a certain context result in Pavlovian association between the Apo-pecking effects [unconditioned stimulus, US] and contextual cues [conditioned stimulus, CS] and sensitization results from Apo-elicited pecking [unconditioned response, UR] plus pecking elicited by the context [conditioned response, CR]). Naturally, there is the possibility that both pharmacological and conditioning mechanisms participate or interact to produce Apo-sensitization. The main objective of this thesis was to assess the adequacy of the conditioning hypothesis to explain this phenomenon. The general procedure was injecting pigeons i.m. with the corresponding solution (Apo, saline or haloperidol) and placing them immediately into a visually distinctive experimental cage (Di-cage), which was a standard metal grid cage like the ones used to lodge the pigeons (Home-cage), although its side and back walls were lined with white panels speckled with dark green dots. The conditioning hypothesis was first explored using a monocular approach, based on the fact that birds have almost completely crossed optic nerves and negligible transfer between brain hemispheres. Pigeons were trained with repeated Apo+Di-cage util asymptotic response with one eye occluded and subsequently tested with Apo+Di-cage with the contralateral eye occluded. If the sensitization disappeared in the test, it could be concluded that it was fully due to associative learning, which failed to be transferred between hemispheres. However, the response was still asymptotic in the test, and two possible explanations remained: either sensitization was due to pharmacological changes induced by the circulating Apo (which reached both brain hemispheres) or it was due to an Apo-context learned association, which could be transferred despite the reduced interhemispheric connection. The conditioning hypothesis was assessed by training pigeons with different combinations of the proposed US and CS (different groups were exposed to the stimuli either consistently paired, explicitly unpaired, only Apo, only Di-cage or none of them). The results suggested that the Apo-induced sensitization was due to a main Apo-context conditioning component plus a minor apparently learning-independent component. In another experiment, pigeons were trained in the Home-cage and tested either in the Di-cage or in a very different Ab-cage. Results showed that they developed an Apo+Home-cage association that generalized to the relatively similar Apo+Di-cage, but not to the rather different Apo+Ab-cage. Thus, when the training and test contexts were markedly different, all traces of a learning-independent component disappeared. This suggests that Apo-sensitization can be fully explained through learning mechanisms involving classical conditioning plus context generalization. The following experiments pursued to further characterize this learning model. The effects of an extinction treatment were explored by injecting saline instead of Apo to trained pigeons until their response sensitization waned and finally extinguished. However, in a subsequent test with Apo+Di-cage the full sensitization re-emerged with response at asymptotic levels. Such resistance to inhibitory conditioning could be due to pigeons perceiving the context differently under Apo than saline. This idea led to an experiment designed to avoid direct Apo vs. no-Apo comparison: two groups were trained with either a larger (1.0 mg/kg) or a smaller (0.2 mg/kg) Apo dose until asymptotic response and then switched for further trials. An analysis of the CR/UR ratios revealed that the Apo dose switch induced a clear positive behavioral contrast and only a slight tendency to negative contrast. It was concluded that the Apo-induced pecking sensitization of pigeons is a due to Pavlovian association between the Apo pecking effects and the context; and that it can be explained through conditioned pecking elicited by the context on top of the unconditioned pecking elicited by Apo, without the involvement of learning-independent processes.