Task-dependent modulation of word processing mechanisms during modified visual search tasks

Abstract

During visual search for words, the impact of the visual and semantic features of words varies as a function of the search task. This event-related potential (ERP) study focused on the way these features of words are used to detect similarities between the distractor words that are glanced at and the target word, as well as to then reject the distractor words. The participants had to search for a target word that was either given literally or defined by a semantic clue among words presented sequentially. The distractor words included words that resembled the target and words that were semantically related to the target. The P2a component was the first component to be modulated by the visual and/or semantic similarity of distractors to the target word, and these modulations varied according to the task. The same held true for the later N300 and N400 components, which confirms that, depending on the task, distinct processing pathways were sensitized through attentional modulation. Hence, the process that matches what is perceived with the target acts during the first 200 ms after word presentation, and both early detection and late rejection processes of words depend on the search task and on the representation of the target stored in memory.

Keywords:

• Categorization
• Electroencephalography
• Event-related potentials
• Cognitive task-set
• Word processing pathways

Notes

¹The response times were also analysed using a full three-way ANOVA with the type of word (orthographic distractor, semantic distractor, or filler word), the task (literal vs. categorical), and the relatedness condition (related vs. unrelated distractors) as within-subjects factors, as in Dampuré et al. (2014). The results confirmed in all respects the statistical analysis presented in the main text. They also showed that the response times obtained for UODs and USDs were not significantly different from those obtained for filler words, indicating that UODs and USDs can be used as unrelated control words. However, for the sake of consistency with the analysis of ERP data, the choice was made to present separate ANOVAs for each type of distractor.