Cue response dissociates inhibitory processes: task identity information is related to backward inhibition but not to competitor rule suppression

Abstract

In task switching, a conflict between competing task-sets is resolved by inhibiting the interfering task-set. Recent models have proposed a framework of the task-set as composed of two hierarchical components: abstract task identity (e.g., respond to quantity) and more concrete task rules (e.g., category-response rules mapping the categories “one” and “three” to the left and right keys, respectively). The present study explored whether task-set inhibition is the outcome of a general control process or whether it reflects multiple inhibitory processes, each targeting a different component of the competing task-set. To this end, two effects of task-set inhibition were examined: backward inhibition (BI), reflecting the suppression of a just-performed task-set that is no longer relevant; and, competitor rule suppression (CRS), reflecting the suppression of an irrelevant task-set that generates a response conflict. In two task switching experiments, each involving three tasks, we asked participants to make two responses: a cue response, indicating the identity of the relevant task (e.g., “Color”), and a target response requiring the implementation of the task rule (e.g., “Red”). The results demonstrate that BI, but not CRS, appears in cue responses, and thus, suggests that BI reflects inhibition that influences representations related to abstract task identity, rather than (just) competing responses or response rules. These results support a dissociation between inhibitory processes in task switching. The current findings also provide further evidence for a multi-component conceptualization of task-set and task-set inhibition.

Appendix: Explaining the orthogonality of BI and CRS

The four combinations of CRS (present or absent, +, −) and BI are presented in the following table:

	CRS−	CRS+
BI−	n − 2 a*b*c*	n − 2 a*b*c*
	n − 1 a1b1c2	n − 1 a2b1c1
	n a*b*c*	n a*b*c*
BI+	n − 2 a*b*c*	n − 2 a*b*c*
	n − 1 a1b1c2	n − 1 a2b1c1
	n a*b*c*	n a*b*c*

In this notation (see Meiran et al., 2010), a1, a2 represent target stimuli associated with Key1 and Key2, respectively, according to Task A. Similarly, b1, b2 and c1, c2 represent target stimuli associated with Key1 and Key2, respectively, according to Tasks B and C, respectively. When the key association is unimportant, the digit (1, 2) is replaced with an asterisk. The currently relevant task dimension is underlined.

Note that what dictates the presence/absence of BI is the task sequence, regardless of task conflict (A-B-A vs. C-B-A). What dictates the presence of CRS is the relationship between the conflict in Trial n − 1 and the relevant task in Trial n. Without loss of generality, the conflicting response in the table is always Key 2, and the correct response is always Key 1. Thus, in the CRS− condition, the conflict (in Trial n − 1) comes from Task C, while in CRS+ trials it comes from Task A.