Limits...
The role of temporal predictability for early attentional adjustments after conflict

View Article: PubMed Central - PubMed

ABSTRACT

A frequently-studied phenomenon in cognitive-control research is conflict adaptation, or the finding that congruency effects are smaller after incongruent trials. Prominent cognitive control accounts suggest that this adaptation effect can be explained by transient conflict-induced modulations of selective attention, reducing congruency effects on the next trial. In the present study, we investigated these possible attentional modulations in four experiments using the Stroop and Flanker tasks, dissociating possible enhancements of task-relevant information from suppression of task-irrelevant information by varying when this information was presented. In two experiments, the irrelevant stimulus information was randomly presented shortly before, at the same time, or briefly after the presentation of the relevant dimension. In the other two, irrelevant information was always presented first, making this aspect fully predictable. Despite the central role that attentional adjustments play in theoretical accounts of conflict adaption, we only found evidence for such processes in one of the four experiments. Specifically, we found a modulation of the attention-related posterior N1 event-related potential component that was consistent with paying less attention to the irrelevant information after incongruent trials. This was accompanied by increased inter-trial mid-frontal theta power and a theta-power conflict adaptation effect. We interpret these results as evidence for an adaptive mechanism based on relative attentional inhibition. Importantly, this mechanism only clearly seems to be implemented in a very specific context of high temporal predictability, and only in the Flanker task.

No MeSH data available.


Related in: MedlinePlus

Results of experiment 3&4—Flanker experiments with predictable and unpredictable timing.(A)(B) Mean reaction times per SOA condition (irrelevant-first, simultaneous and relevant-first in experiment 1, (A); irrelevant-first in experiment 2, (B)) for current congruent (dark grey) and incongruent (light grey) trials as a function of the congruency level of the previous trial in the Flanker experiments (error bars represent 1 standard deviation around the mean). (C)(D) Early attentional ERP component amplitudes (μV) per SOA condition as a function of the congruency level of the previous trial in the Flanker experiments with unpredictable (C) and predictable (D) timing. Measurements were performed for averaged activity over a 50-ms window between 150 and 200 ms post-stimulus onset for the collapsed posterior ROIs, indicated in red (left posterior sensors PO3, P3 and P1, and right posterior sensors PO4, P4 and P2). The irrelevant-first and relevant-first conditions provided an uninterrupted window for 200 ms, during which no other overlapping stimulus response activity would be present. Only the Flanker task with a predictable irrelevant-first temporal arrangement showed a significant posterior modulation reflecting a decreased negativity starting around 150 ms, likely representing decreased early attentional processing.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC5391946&req=5

pone.0175694.g003: Results of experiment 3&4—Flanker experiments with predictable and unpredictable timing.(A)(B) Mean reaction times per SOA condition (irrelevant-first, simultaneous and relevant-first in experiment 1, (A); irrelevant-first in experiment 2, (B)) for current congruent (dark grey) and incongruent (light grey) trials as a function of the congruency level of the previous trial in the Flanker experiments (error bars represent 1 standard deviation around the mean). (C)(D) Early attentional ERP component amplitudes (μV) per SOA condition as a function of the congruency level of the previous trial in the Flanker experiments with unpredictable (C) and predictable (D) timing. Measurements were performed for averaged activity over a 50-ms window between 150 and 200 ms post-stimulus onset for the collapsed posterior ROIs, indicated in red (left posterior sensors PO3, P3 and P1, and right posterior sensors PO4, P4 and P2). The irrelevant-first and relevant-first conditions provided an uninterrupted window for 200 ms, during which no other overlapping stimulus response activity would be present. Only the Flanker task with a predictable irrelevant-first temporal arrangement showed a significant posterior modulation reflecting a decreased negativity starting around 150 ms, likely representing decreased early attentional processing.

Mentions: The main effects of SOA and current congruency were significant, F(1.36,29.91) = 134.51, p < .001, r = .93 and F(1, 21) = 554.01, p < .001, r = .98, respectively, whereas the main effect of previous congruency was not, F(1, 21) = .1, p >.3, r = .20. Participants responded faster to irrelevant-first trials compared to simultaneous and relevant-first trials. There was a significant interaction between SOA and current congruency, F(1.49,32.80) = 221.82, p < .001, r = .95, showing that the congruency effect on trial n was largest for trials in which the distracter preceded the target and smallest for trials in which the target preceded the distracter. The interaction between SOA and previous congruency did not reach significance, F(2, 42) = 1.65, p >.2, r = .26. An overall conflict adaptation effect, as reflected in the interaction between previous congruency and current congruency, was present, F(1,21) = .7.85, p < .05, r = .51. Importantly, the three-way interaction between SOA, previous congruency and current congruency was highly significant, F(1, 42) = 11.23, p < .001, r = .58. The conflict adaptation effect was significant for trials in which the irrelevant word was presented first, F(1,21) = 20.61, p < .001, r = .7: the congruency effect was smaller after incongruent trials (86 ms) than after congruent trials (101 ms). For trials where the relevant and irrelevant dimension were presented at the same time, the conflict adaptation effect was not significant, F(1,21) = 2.50, p >.1, r = .32. Also for the condition in which relevant information was presented first, we could not find a significant conflict adaptation effect, F(1,21) = 1.15, p >.2, r = .22 (Fig 3A and Table 3).


The role of temporal predictability for early attentional adjustments after conflict
Results of experiment 3&4—Flanker experiments with predictable and unpredictable timing.(A)(B) Mean reaction times per SOA condition (irrelevant-first, simultaneous and relevant-first in experiment 1, (A); irrelevant-first in experiment 2, (B)) for current congruent (dark grey) and incongruent (light grey) trials as a function of the congruency level of the previous trial in the Flanker experiments (error bars represent 1 standard deviation around the mean). (C)(D) Early attentional ERP component amplitudes (μV) per SOA condition as a function of the congruency level of the previous trial in the Flanker experiments with unpredictable (C) and predictable (D) timing. Measurements were performed for averaged activity over a 50-ms window between 150 and 200 ms post-stimulus onset for the collapsed posterior ROIs, indicated in red (left posterior sensors PO3, P3 and P1, and right posterior sensors PO4, P4 and P2). The irrelevant-first and relevant-first conditions provided an uninterrupted window for 200 ms, during which no other overlapping stimulus response activity would be present. Only the Flanker task with a predictable irrelevant-first temporal arrangement showed a significant posterior modulation reflecting a decreased negativity starting around 150 ms, likely representing decreased early attentional processing.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC5391946&req=5

pone.0175694.g003: Results of experiment 3&4—Flanker experiments with predictable and unpredictable timing.(A)(B) Mean reaction times per SOA condition (irrelevant-first, simultaneous and relevant-first in experiment 1, (A); irrelevant-first in experiment 2, (B)) for current congruent (dark grey) and incongruent (light grey) trials as a function of the congruency level of the previous trial in the Flanker experiments (error bars represent 1 standard deviation around the mean). (C)(D) Early attentional ERP component amplitudes (μV) per SOA condition as a function of the congruency level of the previous trial in the Flanker experiments with unpredictable (C) and predictable (D) timing. Measurements were performed for averaged activity over a 50-ms window between 150 and 200 ms post-stimulus onset for the collapsed posterior ROIs, indicated in red (left posterior sensors PO3, P3 and P1, and right posterior sensors PO4, P4 and P2). The irrelevant-first and relevant-first conditions provided an uninterrupted window for 200 ms, during which no other overlapping stimulus response activity would be present. Only the Flanker task with a predictable irrelevant-first temporal arrangement showed a significant posterior modulation reflecting a decreased negativity starting around 150 ms, likely representing decreased early attentional processing.
Mentions: The main effects of SOA and current congruency were significant, F(1.36,29.91) = 134.51, p < .001, r = .93 and F(1, 21) = 554.01, p < .001, r = .98, respectively, whereas the main effect of previous congruency was not, F(1, 21) = .1, p >.3, r = .20. Participants responded faster to irrelevant-first trials compared to simultaneous and relevant-first trials. There was a significant interaction between SOA and current congruency, F(1.49,32.80) = 221.82, p < .001, r = .95, showing that the congruency effect on trial n was largest for trials in which the distracter preceded the target and smallest for trials in which the target preceded the distracter. The interaction between SOA and previous congruency did not reach significance, F(2, 42) = 1.65, p >.2, r = .26. An overall conflict adaptation effect, as reflected in the interaction between previous congruency and current congruency, was present, F(1,21) = .7.85, p < .05, r = .51. Importantly, the three-way interaction between SOA, previous congruency and current congruency was highly significant, F(1, 42) = 11.23, p < .001, r = .58. The conflict adaptation effect was significant for trials in which the irrelevant word was presented first, F(1,21) = 20.61, p < .001, r = .7: the congruency effect was smaller after incongruent trials (86 ms) than after congruent trials (101 ms). For trials where the relevant and irrelevant dimension were presented at the same time, the conflict adaptation effect was not significant, F(1,21) = 2.50, p >.1, r = .32. Also for the condition in which relevant information was presented first, we could not find a significant conflict adaptation effect, F(1,21) = 1.15, p >.2, r = .22 (Fig 3A and Table 3).

View Article: PubMed Central - PubMed

ABSTRACT

A frequently-studied phenomenon in cognitive-control research is conflict adaptation, or the finding that congruency effects are smaller after incongruent trials. Prominent cognitive control accounts suggest that this adaptation effect can be explained by transient conflict-induced modulations of selective attention, reducing congruency effects on the next trial. In the present study, we investigated these possible attentional modulations in four experiments using the Stroop and Flanker tasks, dissociating possible enhancements of task-relevant information from suppression of task-irrelevant information by varying when this information was presented. In two experiments, the irrelevant stimulus information was randomly presented shortly before, at the same time, or briefly after the presentation of the relevant dimension. In the other two, irrelevant information was always presented first, making this aspect fully predictable. Despite the central role that attentional adjustments play in theoretical accounts of conflict adaption, we only found evidence for such processes in one of the four experiments. Specifically, we found a modulation of the attention-related posterior N1 event-related potential component that was consistent with paying less attention to the irrelevant information after incongruent trials. This was accompanied by increased inter-trial mid-frontal theta power and a theta-power conflict adaptation effect. We interpret these results as evidence for an adaptive mechanism based on relative attentional inhibition. Importantly, this mechanism only clearly seems to be implemented in a very specific context of high temporal predictability, and only in the Flanker task.

No MeSH data available.


Related in: MedlinePlus