Effects of conflict trial proportion: A comparison of the Eriksen and Simon tasks

Abstract

Two experiments examined global and local behavioral adaptation effects within and across the Eriksen task, where conflict is based on stimulus letter identities, and the Simon task, where conflict is based on stimulus and response locations. Trials of the two tasks were randomly intermixed, and the list-wide proportion of congruent trials was varied in both tasks (Experiment 1) or in just one task (Experiment 2). The global adaptation effect of list-wide congruency proportion (LWPC effect) was at least as large in the Simon task as in the Eriksen task. Likewise, the local adaptation effect of previous-trial congruency (Gratton effect) was at least as large in the Simon task as in the Eriksen task. In contrast to prior studies investigating transfer across Stroop and Simon tasks, there was no dissociation between global and local adaptation effects regarding their transfer across the different conflict tasks. In fact, both local and global adaptation effects appeared largely task-specific, because there was no or only little transfer of either Gratton effects or LWPC effects from the Eriksen to the Simon task or vice versa. On the whole, the results suggest that behavioral adaptation observed in the present design does not carry over from one of these tasks to the other, suggesting no involvement of a higher-order, task-general mechanism of cognitive control.

Keywords: Attention and executive control; Cognitive control; Conflict tasks; Congruency proportion; Contingency learning; Eriksen task; Gratton effect; Simon task.

Fig. 1

Mean congruency effects in Experiment 1 for reaction time (top row) and percentage of correct responses (bottom row) as a function of current task and congruency proportion. The LWPC effect corresponds to the difference between adjacent filled and unfilled bars. The error bars reflect ± 1 within-subjects standard error of the mean, computed using the method of Morey (2008)

Fig. 2

Mean congruency effect in Experiment 1 for reaction time (top row) and percentage of correct responses (bottom row), as a function of current task, previous task, and previous congruency. For each combination of current task and previous task, the Gratton effect is the difference between the adjacent filled and unfilled bars. The error bars reflect ± 1 within-subjects standard error of the mean, computed using the method of Morey (2008)

Fig. 3

Mean congruency effect in Experiment 2 for reaction time (upper panels) and percentage of correct responses (PC, lower panels) as a function of current task, congruency proportion, and inducer task. The LWPC effect corresponds to the difference between adjacent filled and unfilled bars. The error bars reflect ± 1 within-subjects standard error of the mean, computed using the method of Morey (2008)

Fig. 4

Mean congruency effect in Experiment 2 for reaction time as a function of current task, previous task, previous congruency, and inducer task. The Gratton effect is the difference between the adjacent filled and unfilled bars. The error bars reflect ± 1 within-subjects standard error of the mean, computed using the method of Morey (2008)

Fig. 5

Mean congruency effect in Experiment 2 for percentage of correct responses as a function of current task, previous task, previous congruency, and inducer task. The Gratton effect is the difference between the adjacent filled and unfilled bars. The error bars reflect ± 1 within-subjects standard error of the mean, computed using the method of Morey (2008)

Fig. 6

Mean reaction time (RT, in ms, upper panels) and mean percentage of correct responses (PC, lower panels) as a function of congruency, current task, and congruency proportion in Experiment 1. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 7

Mean reaction time (RT, in ms) as a function of congruency, current task, previous task, and previous congruency in Experiment 1. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 8

Mean percentage of correct responses (PC) as a function of congruency, current task, previous task, and previous congruency in Experiment 1. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 9

Mean reaction time (RT, in ms) as a function of congruency, current task, inducer task, and congruency proportion in Experiment 2. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 10

Mean percentage of correct responses (PC) as a function of congruency, current task, inducer task, and congruency proportion in Experiment 2. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 11

Mean reaction time (RT, in ms) as a function of congruency, current task, previous task, previous congruency, and inducer task in Experiment 2. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)

Fig. 12

Mean percentage of correct responses (PC) as a function of congruency, current task, previous task, previous congruency, and inducer task in Experiment 2. Error bars reflect ± 1 within-subjects standard error of the mean according to a suggestion of Morey (2008)