Cortical correlates of selective attention and action selection during decision-making under conflict

Human-beings have to constantly attend to, and act towards, stimuli which are the most compatible with their goals, while refraining from selecting less relevant alternatives. Hence, human behavior emerges from a competition between potential options, which occurs at both the sensory and the motor levels. Two neural processes help the resolution of this competition. First, selective attention biases competitive interactions between sensory stimuli by facilitating neural activity related to the processing of goal-relevant information, while suppressing activity for the processing of irrelevant information. Second, action selection results from the accumulation of activity in competing motor representations and ensures that only activity related to the most beneficial action is ultimately selected. Here, we used electroencephalography (EEG) to characterize the cortical correlates of selective attention and action selection during motor decisions under conflict. Seventeen healthy subjects performed a version of the Flanker task in which they were asked to indicate by a left or right button-press the orientation of a briefly presented left- or right-facing central arrow, flanked by a set of two distractor arrows on each side which either pointed in the same (congruent trials) or in the opposite direction (incongruent trials). The percentage of congruent and incongruent trials was manipulated to produce two different contexts in terms of conflict expectation. In mostly congruent contexts (MCC), 80% of the trials were congruent while in mostly incongruent contexts (MIC), 80% were incongruent. Hence, subjects expected more conflict in MIC than in MCC. Selective attention and action selection processes were assessed by recording steady-state visual evoked potentials (SSVEPs, central and distractor arrows flickered at different frequencies) and response-locked potentials measured over the motor cortex, respectively. At the behavioral level, accuracy was higher and reaction times lower in congruent than in incongruent trials. Interestingly, these differences were accentuated in MCC compared to MIC, which suggests that the level of conflict predictability altered the recruitment of control mechanisms. As such, our EEG results indicates that subjects adopted different strategies depending on context, impacting both selective attention and action selection processes. Most noticeably, subjects focused more on the target and showed a better lateralization of motor cortical activity when incongruent trials occurred in MCC compared to MIC, probably because they could less anticipate the occurrence of conflictual distractor signals in the former than in the latter context. This strategy may allow to limit the decrease of performance when sensory-motor conflict occurs while it is unpredictable