It has been suggested that task-switching costs can be eliminated if participants memorise all stimulus-response mappings thereby avoiding task-switching altogether (Dreisbach, Goschke & Haider, 2006, 2007; Dreisbach & Haider, 2008). This has been labelled the “Look-Up Table” (LUT) approach. It has also been suggested that the LUT approach could potentially explain why animals such as monkeys (Stoet & Snyder, 2003; Avdagic et al., 2013) and pigeons (Castro & Wasserman, 2016; Meier, Lea & McLaren 2016) were able to perform task-switching without showing any task-switching costs (Dreisbach, et al., 2006, 2007; Dreisbach & Haider, 2008; Forrest, Monsell & Mclaren, 2014). In a series of eight experiments the following two questions were addressed: (1) Why do some participants show significant task-switching costs even when they do not switch between tasks (e.g., Forrest, Monsell & Mclaren, 2014)? (2) Can the LUT approach explain the absence of task-switching costs? In an attempt to answer both questions different sources of human task-switching costs are investigated in eight behavioural experiments. Chapter 1 provides an overview of different task-switching paradigms and accounts to explain task-switching costs. Chapter 2 summarises previous attempts to remove human task-switching costs. Evidence for the absence of task-switching costs in animals is also introduced. Following up on previous studies that suggested the LUT approach can explain the absence of task-switching costs, I conducted two task-switching experiments using visual tasks (i.e., colour task and shape task) with bivalent stimuli in an attempt to re-examine the conclusions of previous LUT studies (i.e., Dreisbach, et al., 2006, 2007; Dreisbach & Haider, 2008; Forrest, Monsell & Mclaren, 2014). The results in Chapter 2 indicate that human participants cannot always eliminate task-switching costs and do not always apply the LUT approach when the task-switching strategy is controlled. Therefore, the experiments in Chapter 3 and 4 sought to ascertain the requirements for eliminating task-switching costs when using the LUT approach. The experiments in Chapter 3 applied visual tasks where each task had a different stimulus-set. Experiments in Chapter 4 applied two classical mathematical tasks (i.e., big/small task, odd/even task) and used Chinese numbers as stimuli. The results of the experiments in Chapters 3 and 4 suggest that human participants must be able to give the correct answer without processing task-relevant features from the stimuli in order to eliminate task-switching costs. In the experiment of Chapter 5 the cue-stimulus-response mappings from Experiments 2.1 and 2.2 were rearranged so that switching between conventional tasks and rules became impossible. The results suggest that task-relevant features can trigger interferences thereby causing “task-switching costs” even when participants do not switch between tasks. In Chapter 7, I compare a modified interference account, introduced in Chapter 5, with the compound retrieval account (e.g., Logan & Schneider, 2010) and associative learning account (Forrest et al., 2014; Meier et al., 2016) in order to explain why human participants show task-switching costs even when they do not switch between tasks. I conclude that the modified interference account provides an alternative explanation. It has been proposed that only humans are affected by strong and long-lasting interference from previous trials during task-switching. As a consequence, this interference may explain why human participants consistently show task-switching costs whereas monkeys and pigeons show no task-switching costs.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:739283 |
| Date | January 2018 |
| Creators | Li, Xiangqian |
| Publisher | University of Glasgow |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://theses.gla.ac.uk/8962/ |
Page generated in 0.0022 seconds