The role of temporal order of relevant and irrelevant dimensions within conflict tasks.

Abstract

Conflict tasks are commonly used to investigate control processes under situations of relevant and irrelevant sources of information. In addition to compatibility effects at a mean behavioral level, delta plot analyses of reaction time distributions reveal that the compatibility effect generally increases with time (i.e., positive delta plot slopes) across most conflict-like tasks. Critically, the underlying causes of the increasing delta plot slopes with different types of distractors are still poorly understood. The present study tested whether the relative onset of target-to-distractor processing affects the delta plot slope. Specifically, we manipulated the temporal order of relevant and irrelevant dimensions within an Eriksen flanker task (Experiment [Exp.] 1), an Arrow-Simon task (Exp. 2), and a manual Stroop task (Exp. 3a/3b). The results of the Eriksen flanker task and Arrow-Simon task revealed that the delta plots slopes were less increasing (and instead rather decreasing) when the irrelevant dimension appears first (IR condition) compared to the reversed order (RI condition)-consistent with the idea that the underlying mechanism driving the slope of the delta plot is the temporal overlap of activation between the relevant and irrelevant dimensions. In contrast, for the Stroop task, the delta plots in the RI condition were not more increasing than the ones for the IR condition. Overall, these results suggest that the temporal properties strongly influence delta plot shape, but that the temporal dynamics operating in the flanker task and the Arrow-Simon task differs from the Stroop task, at least under conditions where relevant and irrelevant information is presented sequentially. Public Significance StatementIn many real-world situations, people are required to select and process goal-relevant information in environments overloaded with distracting-and potentially conflicting-sources of information. It is important to uncover the mental control mechanisms that prevent interference allowing successful goaldirected behavior for both practical and theoretical reasons. In the present study, we systematically investigated the temporal processing dynamics with different sources of distracting information. The results indicate that some types of distracting information are similarly processed across time whereas others are not, suggesting the existence of distractor-general and distractor-specific control mechanisms.

Staggered Onsets of Processing Relevant and Irrelevant Stimulus Features Produce Different Dynamics of Congruency Effects

“…This characteristic of the delta plots is consistent with the hypothesis of a lag of processing the task-irrelevant feature relative to processing the task-relevant feature, as indicated by the estimated mean temporal offset of the LCA models. A similar shape of delta plots has also been observed when irrelevant features were presented after the relevant ones in a Simon task or an Eriksen flanker task ( Burle et al, 2005 ; Mackenzie et al, 2022 ). Our next step is fitting the models to a Simon task, for which we hypothesize a lead rather than a lag of processing the task-irrelevant stimulus feature when both features are presented simultaneously.…”

“…In contrast, when the irrelevant feature is presented after the relevant one, delta plots increase with sometimes a delayed appearance of congruency effects (e.g., Burle et al, 2005 ; Mackenzie et al, 2022 ). An exception from this pattern has been found for the Stroop task, suggesting that other factors than temporal overlap – and delays of processing in particular – also modulate the shape of delta plots ( Mackenzie et al, 2022 ).…”

“…Reaction time was measured from the onset of the task-relevant stimulus feature (green or red color of the square) or the onset of the task-irrelevant feature (lateral position of the square), whatever was first, until the keypress. In this way, we treat the experimentally defined temporal offset as if it were an internally generated offset, which differs from other studies in which presentation of the relevant feature triggered reaction-time measurements (e.g., Burle et al, 2005 ; Mackenzie et al, 2022 ). As the only consequence of this difference, reaction times for negative SOAs were longer than at SOAs ≥ 0 in the present experiments rather than shorter as when the SOA is not a part of the reaction time.…”

“…1 From the bin means of each pair of congruent and incongruent conditions, we computed the delta plot for each participant and SOA and characterized its overall increase or decrease across reaction-time bins by the slope of a linear regression (cf. Mackenzie et al, 2022 ). For these dependent variables, we report the results of statistical analyses.…”

“…The variety of shapes of delta plots does not only call for an explanation of decreasing delta plots which violate the typical relation between reaction-time means and variances, but also for an explanation of the very existence of different shapes. Such accounts are of basically two types, namely in terms of qualitative differences such as different processes or codes involved (e.g., De Jong et al, 1994 ; Gade, Paelecke, & Rey-Mermet, 2020 ; Wiegand & Wascher, 2005 ; Wühr & Biebl, 2011 ) or in terms of quantitative differences such as variations of temporal overlap of processing relevant and irrelevant stimulus features (e.g., Mackenzie, Mittelstädt, Ulrich, & Leuthold, 2022 ). Although these two claims can be contrasted (e.g., Töbel, Hübner, & Stürmer, 2014 ), qualitative differences such as the involvement of spatial codes in the horizontal but verbal codes in the vertical Simon task ( Gade et al, 2020 ; Wühr & Biebl, 2011 ) will likely go along with quantitative differences of the temporal overlap of processing relevant and irrelevant features.…”