Academic journal article Canadian Journal of Experimental Psychology

The Target of Task Switching

Academic journal article Canadian Journal of Experimental Psychology

The Target of Task Switching

Article excerpt

Task switching involves processing target stimuli in accordance with a frequently changing series of tasks. An outstanding issue is whether this processing is tailored to the perceptual or categorical representation of targets. To address this issue, the authors compared switch costs in responding to targets that were perceptually distinct (words and images) but associated with the same categories (colors and shapes). In four experiments that varied the degree to which words and images were mixed together, no differences in switch costs were observed. These results support the idea that categorical target representations are central to task switching.

Keywords: task switching, switch cost, categorical representation, target encoding

Task switching involves processing target stimuli in accordance with a frequently changing series of tasks. How this feat is accomplished is a matter of debate, but there are models of task switching (e.g., Altmann & Gray, 2008; Gilbert & Shallice, 2002; Meiran, Kessler, & Adi-Japha, 2008; Schneider & Logan, 2005, 2009; Sohn & Anderson, 2001) that specify not only how task switching occurs, but also how tasks are performed (i.e., how targets are processed to generate responses).

An important issue concerns the level at which targets should be represented in a model. In this article we distinguish between two levels of representation: perceptual and categorical. A perceptual representation is one in which the model codes for the perceptual attributes of the target. For example, the perceptual representation of a colored shape target would be coded in terms of chromatic and geometric properties. A categorical representation is one in which the model codes for the task categories associated with the target. For example, the categorical representation of a colored shape target would be coded in terms of color and shape categories (e.g., blue and circle).

Existing models of task switching assume that targets are represented categorically (e.g., Altmann & Gray, 2008; Gilbert & Shallice, 2002; Meiran et al., 2008; Schneider & Logan, 2005, 2009; Sohn & Anderson, 2001). This assumption is a virtue in that it allows these models to be applied to different target types. However, this virtue may be a vice if task switching is actually tailored to the perceptual representation of targets. In this case, models would lose considerable generality because they would have to be specialized for different target types. Thus, determining the nature of the target representation involved in task switching is important for assessing the generality of models.

How can one distinguish between perceptual and categorical target representations? One way is to compare task-switching performance for target types that are perceptually distinct but associated with the same categories. The performance measure used in the present study was switch cost - slower performance for task switches than for task repetitions. We compared switch costs when switching between color (red, green, or blue) and shape (circle, triangle, or square) tasks for two target types: words and images. Examples of each target type are shown in Figure 1 . Word targets were pairs of words denoting colors and shapes (e.g., red triangle). Image targets were irregularly shaped color patches beside white-outlined shapes. Each task was signaled by one of two cues, enabling three types of task transitions across trials: task switches (cue and task both switch), task repetitions (cue switches but task repeats), and cue repetitions (cue and task both repeat). We define switch cost as the difference between task switches and task repetitions to avoid confounding task switching with cue switching (Logan & Bundesen, 2003; Schneider & Logan, 2005, 2007).

We predicted different patterns of switch costs depending on whether targets are represented perceptually or categorically for the purposes of task-switching performance. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.