A Comparative Investigation of Associative Processes in Executive-Control Paradigms

Abstract

The experiments reported in this thesis were conducted to examine the effects of executive-control and associative-learning processes on performance in conventional executive-control paradigms. For this purpose, I developed comparative task-switching and response-inhibition paradigms, which were used to assess the performance of pigeons, whose behaviour is presumably based purely on associative processes, and of humans, whose behaviour may be guided by executive control and by associative processes. Pigeons were able to perform accurately in the comparative paradigms; hence, associative-learning processes are sufficient to account for successful performance. However, some task-specific effects that can be attributed to executive-control processes, and which were found in humans applying executive control, were absent or greatly reduced in pigeons. Those effects either reflect the mental operations that are performed to ensure that a specific set of stimulus-response-contingencies is applied and any contingencies belonging to a different set are suppressed, or reflect mental preparations for the possibility that the requirement to execute a certain response suddenly changes. In particular, in Chapter 3, it is shown that the benefits of repeatedly applying the same set of stimulus-response contingencies (or, in reverse, the costs of switching from one set to another) do not apply when Pavlovian processes dominate learning, which is likely the case for pigeons. Furthermore, as shown in Chapters 4 and 5, the behavioural effects of preparing for an unpredicted change in response requirements appeared to be absent when behaviour was based purely on associative processes. Instead, associatively mediated performance was primarily influenced by the stimulus-response contingencies that were effective in each paradigm. Repeating the same response in consecutive trials facilitated the performance of pigeons and associatively learning human participants in the task-switching paradigms, and performing a particular Go response increased the pigeons' likelihood of executing that response in the following trial in two response-inhibition paradigms. In summary, any behavioural effects that can be observed at the level of abstract task requirements reflect the influence of executive-control processes, both in task-switching paradigms and in response-inhibition paradigms.