Reaching into Cluttered Visual Environments: Evidence for Action Representations of Non-target Objects
Among the most important behavioural abilities possessed by humans and other animals is the ability to act selectively in complex sensory environments that afford a large number of potential actions. A common example of such selective action used by experimental psychologists is picking a ripe (red) apple among unripe (non-red) apples. In the laboratory, this selection problem is usually examined using colour visual search tasks, in which subjects have to make a spatially-arbitrary response when they detect a red target among yellow non-targets. Typically selective processes handle such a selection problem very efficiently - measured by no increase in response time when non-targets are present. But such search tasks are poor analogues of the apple picking problem, because in order to pick the red apple one must direct a reaching response to its spatial location. The experiments in this thesis instead used a selective reaching task in which subjects responded to a red target, in the presence of a yellow non-target, by reaching to its location. In the selective reaching task, unlike in the search task, selective processes do not handle the selection problem with optimal efficiency - measured by an increase in response time when non-targets are present (i.e., interference). The experiments in the thesis were designed to show that this suboptimal selection efficiency was related to the fact that reaching responses were used. The hypothesis was informed by physiological evidence that the visuomotor information and colour information are processed by parallel pathways in the visual system. The need for visuomotor processing leading to reaching distinguishes the selective reaching task from the search task. The experiments provided support for the hypothesis that this visuomotor processing led to the concurrent activation of subovert reaching responses associated with the target and non-target, creating a response competition. The results have implications for theories of selection that purportedly apply to real-world situations such as picking a ripe among unripe apples, but that do not take into account the act of picking the apple itself. Applications of the results to real-world situations in which people reach for the wrong object in multiple-object visual displays are also discussed.