Task goals modulate the activation of part-based versus object-based representations in visual working memory

ABSTRACT

Representations of visual objects in working memory (WM) can be part-based or object-based, and we investigated whether this is determined by top-down control processes. Lateralised change detection tasks were employed where sample objects on one task-relevant side had to be memorized. Contralateral delay activity (CDA) components were measured during the retention period as electrophysiological markers of WM maintenance processes. In two critical task conditions, sample displays contained objects composed of two vertically aligned shapes. In the Parts task, test displays contained a single shape that had to be matched with either of the two sample shapes, encouraging the storage of part-based WM representations. In the Whole task, compound-shape objects shown at test had to be matched with memorized compound objects, which should facilitate the formation of object-based integrated WM representations. CDA amplitudes were significantly larger in the Parts task than in the Whole task, indicative of differences in effective WM load. This suggests that the two individual shapes were represented separately in the Parts task, whereas a single compound object was maintained in the Whole task. These results provide new evidence that changes in task goals can result in qualitative differences in the way that identical visual stimuli are represented in WM.

KEYWORDS:

• Working memory
• top-down control
• event-related brain potentials
• part-based representations
• object-based representations

Acknowledgments

This research was supported by Grant ES/L016400/1 from the Economic and Social Research Council (ESRC), UK, and by Grant RPG-2015-370 from the Leverhulme Trust, UK.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ As can be seen in Figure 4, CDA amplitudes in the two baseline tasks were generally smaller than in the corresponding other task. This was the case both for the Single Baseline relative to the Whole task (t(15) = 4.68, p < .001, d = 0.98), and for the Double Baseline relative to the Parts task (t(15) = 3.89, p = .001, d = 0.50).

Additional information

Funding

This work was supported by the Economic and Social Research Council [ES/L016400/1]; Leverhulme Trust [RPG-2015-370].