Abstract:

Animals use many strategies when navigating through space in search of nourishment and shelter. Though much research has been conducted in order to identify these strategies, there is still little consensus regarding whether they are supported by a general mechanism of learning or by a set of domain-specific mechanisms for spatial navigation. This issue has driven researchers to develop new apparatuses that balance ecological validity (e.g., approximating natural foraging conditions) with a reliable and controlled measure of spatial behavior. A series of experiments established the reliability of a new automated open-field to record behavior, display stimuli, and reinforce behavior in pigeons trained in basic conditioning and cue competition procedures. Pigeons were instrumentally shaped to peck at a single ARENA module when lit. Pigeons then successfully learned to approach and peck a color paired with reinforcement and avoid pecking at a module displaying a color never paired with reinforcement. Pigeons were also trained to use the color of a module as a signal for the location of reinforced module. After learning this conditional discrimination, the color-location assignments were reversed and pigeons rapidly learned the new contingencies. After verifying that color can be use a signal for a reinforced location, colors were used as landmarks that signaled the location of a hidden goal. In separate experiments, cue competition effects (e.g., blocking and overshadowing) common to Pavlovian and instrumental conditioning were also obtained in landmark-based search tasks in both an automated open-field and touchscreen preparation. In an overshadowing procedure, the relative proximity of a landmark to a hidden goal determined the strength of spatial control by the landmark at test. A blocking procedure found poor spatial control by a landmark if it was trained in conjunction with another landmark previously established as a spatial signal for the hidden goal. A deficit in spatial control was not found if the same landmark was trained in compound with a novel landmark and the hidden goal. The parallels between the cue competition effects observed in these landmark-based search tasks and those of conventional associative procedures support a general associative mechanism for navigation.